Characterization of a nitric oxide synthase from the plant kingdom: NO generation from the green alga Ostreococcus tauri is light irradiance and growth phase dependent.

This study characterizes a nitric oxide synthase (NOS) from Ostreococcus tauri, a marine green alga. Like animal NOS sequences, O. tauri NOS contains an oxygenase and reductase domain. The recombinant protein is a functional enzyme. NO production in O. tauri cell cultures is induced in the presence of the NOS substrate l-Arg and is light irradiance and growth phase dependent. The search for a nitric oxide synthase (NOS) sequence in the plant kingdom yielded two sequences from the recently published genomes of two green algae species of the Ostreococcus genus, O. tauri and O. lucimarinus. In this study, we characterized the sequence, protein structure, phylogeny, biochemistry, and expression of NOS from O. tauri. The amino acid sequence of O. tauri NOS was found to be 45% similar to that of human NOS. Folding assignment methods showed that O. tauri NOS can fold as the human endothelial NOS isoform. Phylogenetic analysis revealed that O. tauri NOS clusters together with putative NOS sequences of a Synechoccocus sp strain and Physarum polycephalum. This cluster appears as an outgroup of NOS representatives from metazoa. Purified recombinant O. tauri NOS has a Km for the substrate l-Arg of 12 ± 5 μM. Escherichia coli cells expressing recombinant O. tauri NOS have increased levels of NO and cell viability. O. tauri cultures in the exponential growth phase produce 3-fold more NOS-dependent NO than do those in the stationary phase. In O. tauri, NO production increases in high intensity light irradiation and upon addition of l-Arg, suggesting a link between NOS activity and microalgal physiology.


INTRODUCTION
Nitric oxide (NO) is a ubiquitous intra-and intercellular messenger that functions in many physiological processes in all kingdoms of life. In animals, NO is produced by the enzyme nitric oxide synthase (NOS; EC 1.14.13.39). NOS catalyzes the formation of NO and citrulline from L-Arg in a reaction that requires NADPH as an electron donor and O 2 as a cosubstrate. Three NOS isoforms have been described in animals: the constitutive neuronal NOS (nNOS), endothelial NOS (eNOS), and the inducible NOS (iNOS) (Alderton et al., 2001). NOS is a bimodal enzyme, comprising an N-terminal oxygenase domain (NOSoxy) that binds protoporphyrin IX (heme), 6R-tetra-hydrobiopterin (H 4 B), L-Arg, and a C-terminal reductase domain (NOSred) that binds the cofactors flavin mononucleotide (FMN), flavin adenine dinucleotide (FAD), and NADPH. The two domains are connected by a calmodulin (CaM) binding sequence. The activity of the two constitutive NOS isoforms is dependent on changes in the concentration of intracellular Ca 2+ , whereas iNOS is Ca 2+ -CaM independent (Griffith and Stuehr, 1995).
Genomic and functional analyses indicate that NOS enzymes are present in organisms ranging from bacteria to humans (Gorren and Mayer, 2007). Gram-positive bacteria encode smaller NOS proteins, containing only the oxygenase domain. Bacterial NOS uses nonspecific cellular reductases to produce NO (Wang et al., 2007;Gusarov et al., 2008). In higher plants, there are at least two enzymatic sources of NO production: (1) nitrate reductase, which reduces nitrate to nitrite, and then nitrite to NO (Yamasaki et al., 1999); and (2) a NOS-like enzymatic activity (Corpas et al., 2009). Even though the existence of a plant NOS remains a matter of debate, it has been inferred from measurements of NOS activity in several plant extracts (Cueto et al., 1996;Barroso et al., 1999;Caro and Puntarulo, 1999;Ribeiro et al., 1999;Corpas et al., 2009) and inhibition of NO production using mammalian NOS inhibitors (Cueto et al., 1996;Bright et al., 2006;Valderrama et al., 2007). A recent article focused on the relevance of NOS-derived NO metabolism in plant plastids (Gas et al., 2009). However, no gene or protein with sequence similarity to animal or bacterial NOSs has been reported for plants.
The green alga Ostreococcus tauri is the smallest-known freeliving eukaryote and is an abundant picoeukaryotic group throughout the oceanic zone. This organism belongs to the Prasinophyceae (Chlorophyta), a primitive class within the green plant lineage that evolved after the endosymbiosis event that gave rise to photosynthetic eukaryotes (Díez et al., 2001). The recent sequencing of the O. tauri genome revealed that it has one of the smallest (12.56 Mb) and most compact nuclear genomes known to date (Derelle et al., 2006). This single-celled alga is of particular interest because it shares a common ancestor with higher plants and is considered to be an early-diverging class within the green plant lineage. Studying the genes that algae and higher plants have in common and analyzing how they have evolved and are regulated could provide insight into their functions. These features make O. tauri a potentially powerful biological model to study fundamental cellular processes and gene evolution in photosynthetic eukaryotes.
In this work, we report the functional characterization of the NOS enzyme from O. tauri by heterologous expression in Escherichia coli, prediction of its structural arrangement, and determination of its kinetic parameters and spectroscopic properties. Bacteria carrying the NOS gene display enhanced NO production and cell viability. O. tauri expresses NOS throughout its life cycle, but NO production is highest at the exponential growth phase and increases during high intensity light irradiance. Thus, this report describes a NOS enzyme from the plant kingdom that has most of the characteristics ascribed to animal NOS.

Sequence and Structural Analysis of O. tauri NOS
The NOS sequence was obtained from the published genome sequence of O. tauri. The amino acid sequence was predicted and annotated as a NOS enzyme (Derelle et al., 2006). Interpro and PFAM searches were used to identify conserved domains in the primary sequence, and the NOS was found to contain the characteristic NOSoxy and NOSred domains (Sheta et al., 1994;Lowe et al., 1996). These findings were also supported by the fold assignment results using HHpred and FFAS03, where a NOSoxy (PDB code 1d0c with a score of 289.00; a score of less than 29.5 is significant) and a NOSred domain (PDB code 1tll with a score of 2132.00) were found. Figure 1A shows the O. tauri NOS (Ot NOS) sequence alignment with human eNOS, iNOS, and nNOS. The similarity between segments of aligned Ot NOS and human eNOS that overlap is 44%, whereas that between Ot NOS and both iNOS and nNOS is 45%. Full-length sequence similarity is 41.6, 42.7, and 34.3% for eNOS, iNOS, and nNOS, respectively. Furthermore, binding regions of the heme group and cofactors H 4 B, FMN, FAD, and NADPH are largely conserved among all sequences ( Figures 1A and 1B). The first 90 residues are divergent with respect to other NOS enzymes, except for a putative zinc binding motif. O. tauri NOS contains a C-(x) 3 -C motif (where C is Cys and x any residue) instead of the C-(x) 4 -C zinc binding motif commonly found in animal NOS enzymes ( Figure 1A; Alderton et al., 2001). Although the putative CaM binding site is not well conserved in O. tauri NOS, the CaM binding region seems to correspond to the classical Ca 2+ -dependent 1-5-8-14 motif, based on the positions of conserved hydrophobic residues (Rhoads and Friedberg, 1997). Furthermore, an analysis using HHPred (Sö ding et al., 2005) established that this region is similar to the CaM binding region of nNOS from Gallus gallus (PDB code 2o60, E-value 0.0036). The regions involved in the dimerization interface of NOS proteins (i.e., regions I, II, III, and IV) (Bird et al., 2002) are also conserved ( Figure 1A).
Structural models based on best templates from FFAS03 and HHPred results were built using MODELLER. Figures 2A and 2B show the estimated three-dimensional structure of the O. tauri NOSoxy domain based on the coordinates of Bos taurus eNOSoxy. The model was assessed for quality with PROSA II (Wiederstein and Sippl, 2007) and resulted in a z-score of 29.06, which is similar to that of B. taurus eNOS (z-score 29.36). The cofactor and substrate binding sites in the O. tauri NOSoxy and B. taurus eNOSoxy structures are almost identical (Figures 2A and 2B,insets). Furthermore, the residues that interact with the substrate L-Arg, the cofactor H 4 B, and heme are fully conserved in the O. tauri NOS sequence ( Figure 2C). NOS belongs to the family of Cys-coordinated heme proteins in which the proximal ligand to the heme-Fe is the sulfur atom of an intrinsic Cys residue (Rousseau et al., 2005). In O. tauri NOS, heme is stabilized by bonding to the sulfur atom of Cys-125 ( Figure 2C). The substrate L-Arg above the heme iron atom is stabilized by a H-bonding network involving Trp-301, Tyr-302, and Glu-306 that interact with the terminal nitrogen atom, terminal oxygen atom, and guanidium nitrogen, respectively ( Figure 2C). L-Arg and H 4 B are linked together by a H-bond that is mediated by one of the two propionate groups of the heme. H 4 B is further stabilized by an interaction with the oxygen of the Trp-392 residue ( Figure 2C).
The predicted structure of O. tauri NOSred is also conserved among NOSred domains from animals. Figure 2D shows the superposition of O. tauri NOSred and Rattus norvegicus nNOSred and reveals two main differences between the domains. Ot NOS does not contain the autoinhibitory control element (ACE) that blocks the electron flow in the absence of CaM ( Figure 1A; Salerno et al., 1997). The C2DA loop contributes (A) The sequences of Ot NOS and human eNOS, nNOS, and iNOS were aligned using ClustalX and Genedoc software. Black boxes indicate conserved residues in all four sequences, dark-gray boxes represent conserved residues in three sequences, and light-gray boxes represent conserved residues in two sequences. Amino acids that share no similarity are unshaded. Putative cofactor binding sites for zinc, Heme, H 4 B, CaM, FMN, FAD pyrophosphate, FAD isoalloxazine, NADPH ribose, NADPH adenine, and the C-terminal domain of NADPH are shown. Regions involved in the dimerization interface (i.e., regions I, II, III, and IV) are boxed. (B) Comparison of the domain architecture of NOS from different sources. Eukaryotic NOS enzymes have a Zn binding region, a NOSoxy domain that binds heme, L-Arg, and H 4 B, a CaM binding region, and a NOSred domain with subdomains that bind FMN, FAD, and NAD. Bacterial NOS enzymes contain only a NOSoxy domain. The question mark indicates that the Zn binding motif is not completely conserved. H 4 B/THF means that is not clear whether NOS-containing bacteria synthesize H 4 B, although they can use tetrahydrofolate (THF) as a pterin cofactor (Crane et al., 2010). to the Ca 2+ dependence of CaM-bound activity (Knudsen et al., 2003) and is shorter in O. tauri NOSred than in nNOSred (Figure 2D).
Phylogenetic Analysis of O. tauri NOS BLAST searches against the protein nonredundant database from GenBank were used to collect close putative homologs of O. tauri NOS. The sequences retrieved mainly belonged to the superkingdom Eukaryota and to the Metazoa kingdom. Few exceptions were found, such as a NOS from Amoebozoa (Physarum polycephalum), from bacteria (Synechococcus sp), and from the close homolog Ostreococcus lucimarinus. Figure 3 (see Supplemental Data Set 1 online) shows the majority-rule consensus tree obtained from 1000 replicates of maximum likelihood trees. The numbers near the internal nodes in the figure indicate the relative support resulting from nonparametric bootstrapping. Nodes with bootstrap values of below 40% were collapsed. The tree can be divided into two main divisions. A major cluster contains the three NOS isoforms from vertebrate organisms (in red in Figure 3). The clustering pattern within this group represents functional diversification of NOS types (eNOS, nNOS, and iNOS) ( Figure 3). The second populated cluster contains NOS from invertebrates (phyla Placozoa, Cnidaria, Arthropoda, and Mollusca). In this cluster, we also found representatives from green algae (O. tauri and O. lucimarinus), bacteria (Synechococcus sp), and Amoebozoa (P. polycephalum), which are well established outgroups of the Metazoa kingdom (D) Structural alignment between model of Ot NOSred (green) and nNOSred from R. norvegicus (red). Ot NOS contains a shorter CD2A loop and lacks the ACE segment. Note that the nNOS structure represented lacks part of the ACE segment. Due to the high motility of the ACE segment, it was not modeled; therefore, residues Ser-849 to Gly-872 are not depicted. The first and last residues Ser-849 and Gly-872 of the crystal structure of nNOS are shown in ball and stick in light green. (Adoutte et al., 2000). While the clustering pattern among these organisms is well supported, their insertion in the cluster containing other metazoan representatives (in blue in Figure 3) is not adequately supported, making it difficult to establish deep evolutionary relationships among the NOS enzymes within the cluster.

Expression and Purification of Recombinant O. tauri NOS
Recombinant O. tauri NOS expressed in E. coli appeared as a 119-kD band on an SDS-PAGE gel ( Figure 4A, lane 2), in agreement with the molecular mass expected from the amino acid sequence. E. coli cells transformed with the empty vector did not contain proteins in this mass region ( Figure 4A, lane 1). NOS was purified using 2959-ADP agarose column chromatography. The purified protein migrated as a single band on an SDS-PAGE gel ( Figure 4A, lane 3) and exhibited immunoreactivity with anti-OtNOS antibody ( Figure 4B). Under the conditions described in Methods, 1 to 2 mg of recombinant NOS protein was purified from 1 liter of transgenic E. coli Bl21 cell culture expressing pET24b-OtNOS.
Changes in the heme pocket architecture of purified recombinant protein after the addition of different ligands were monitored by the spectral perturbation effect. Spectra exhibit a broad peak at l 396 nm ( Figure 4C), which is typical of the presence of a high-spin heme prosthetic group. The addition of 1 mM imidazole resulted in the complete conversion of the heme group to a low-spin state, characterized by a peak at l 487 nm. After the formation of the imidazole-NOS complex, the addition of L-Arg resulted in a shift of absorbance maximum at l 410 nm, supporting the reconversion of the heme iron to the high-spin state ( Figure 4C).

Catalytic Activity of Recombinant O. tauri NOS
The activity of purified recombinant NOS was assessed by two methods: (1) NO production by the oxyhemoglobin capture assay and (2) [ 3 H]L-citrulline formation (Table 1). Using the oxyhemoglobin assay, we observed a prolonged burst phase of NO release (;150 s at 0.49 6 0.03 mM NO min 21 ) in reactions containing L-Arg, the cofactor H 4 B, and CaM, followed by an extended period of slower NO release (33% of the initial phase). Similar results were obtained with the [ 3 H]L-citrulline method (0.53 6 0.05 mM citrulline min 21 ) in reactions containing L-Arg, the cofactor H 4 B, CaM, and Ot NOS. L-citrulline formation was confirmed by thin layer chromatography (see Supplemental Figure 1 online). Activity assays indicate that recombinant O. tauri NOS is replete with FAD and FMN, since the enzyme retained its activity without adding these cofactors to the reaction (Table 1), as previously reported for recombinant bovine eNOS (Martasek et al., 1996). The K m of O. tauri NOS for the substrate L-Arg estimated from the double reciprocal plot (see Supplemental Figure 2 online) using the oxyhemoglobin method was 12 6 5 mM. In absence of CaM, NOS retained up to 70% of its activity (Table 1). NO and L-citrulline formation were undetectable in reactions that lacked H 4 B and was inhibited by 80% by the inactive L-Arg analog, L-nitro arginine methylester (L-NAME) ( Table 1). Results obtained from assays using a commercial recombinant iNOS from mouse were comparable to those obtained for O. tauri NOS either with the oxyhemoglobin method or [ 3 H]L-citrulline assay ( Table 1).
The rate of NADPH oxidation by O. tauri NOS was also measured. A basal oxidation of NADPH in the absence of the substrate L-Arg (0.42 6 0.2 min 21 ) was subtracted in all measurements. A similar basal oxidation of NADPH has been reported for measurements of nNOS and eNOS activity (Heinzel et al., 1992;Martasek et al., 1996). In a reaction containing L-Arg and all cofactors, 0.5 mM O. tauri NOS oxidized 1.36 6 0.18 mM NADPH min 21 . NADPH oxidation by O. tauri NOS was completely blocked in the absence of H 4 B or by the addition of the inhibitor L-NAME.

Recombinant O. tauri NOS-Dependent NO Production in Bacteria
The ability of E. coli cell cultures transformed with O. tauri NOS to generate NO was assayed. The substrate L-Arg was added at the time of induction by isopropyl thiogalactoside (IPTG), and NO production was monitored over a 120-min period. E. coli transformed with the empty vector pET24b or the recombinant protein O. tauri NOS did not produce a significant amount of NO under basal conditions ( Figures 5A and 5B). The addition of L-Arg to cells transformed with the empty vector triggered negligible NO production compared with the control. However, bacteria expressing recombinant O. tauri NOS produced up to 2.5-fold more NO than did control cells ( Figures 5A and 5B). The inactive enantiomer D-Arg failed to trigger NO formation in both cultures ( Figure 5B). Gusarov and Nudler (2005) showed that NO confers cytoprotection to oxidative stress in bacteria. To investigate the tolerance of O. tauri NOS-transformed bacteria to oxidative stress, we analyzed the induction of cell death by exposure to H 2 O 2 . Figure 5C shows that, under resting conditions, the fold increase in cell death was 0.6 in NOS-transformed cells compared with those transformed with pET24b alone. H 2 O 2 treatment induced an increase in cell death of 30% in the control strain and was 80% higher than in the H 2 O 2 -treated NOS-transformed cells ( Figure  5C). The ratio of cell death in H 2 O 2 -treated versus untreated cells was similar in both pET24b and NOS-transformed cells, indicating that NOS protection of E. coli from cell death is H 2 O 2 independent. As shown in Figure 5D, cells transformed with NOS had greater NO levels than cells transformed with the empty vector. Even though higher NO levels were measured in NOStransformed bacteria, both strains accumulated NO when exposed to H 2 O 2 ( Figure 5D).

NO Production in O. tauri Cell Cultures
To analyze whether O. tauri produces NO in vivo, cell suspensions were incubated with the NO-specific fluorophore, 4-amino-5-methylamino-2',7'-difluorofluorescein diacetate (DAF-FM DA). NO generation was measured during the 60-min period following  Figure 6A shows that, within the first 30 min, L-Arg promoted NO production in a dose-dependent manner (0.5 to 5 mM L-Arg), reaching maximum production at 5 mM L-Arg (8fold induction). The NOS inhibitors L-NAME, G-monomethyl-Larginine (L-NMMA), and NG-nitro-L-arginine (L-NNA) were used to confirm the source of NO production in L-Arg-treated O. tauri cells. Figure 6B shows that NOS inhibitors and the NO scavenger 2-(4-carboxyphenyl)-4,4,5,5-tetramethylimidazoline-1-oxyl-3oxide decreased NO production. Furthermore, the CaM antagonist trifluoperazine dihydrochloride (TFP) caused a 30% reduction in NO production, and this result correlated with the activity of purified NOS protein in the absence of CaM (Table 1). Microscopy analysis showed that NO fluorescence (excitation 495 nm; emission 515 nm) increased significantly when cells were treated with 5 mM L-Arg compared with the untreated control ( Figure 6C). O. tauri cells also displayed red fluorescence, which is attributed to chlorophyll autofluorescence ( Figure 6C).
NOS activity in O. tauri cells depends on the phase of culture growth and on light irradiance conditions. Figure 7A shows that NO production was highest at the exponential growth phase and rapidly dropped at the start of the stationary growth phase. Shifting the cell culture from a light irradiance of 40 mmol m 22 s 21 to 100 mmol m 22 s 21 triggered an increase in NO production, both in the presence and absence of L-Arg ( Figure 7B). Immunoblot analysis using anti-iNOS indicated that the addition of L-Arg did not increase the level of O. tauri NOS in O. tauri cell culture, whereas the shift from low to high irradiance induced the accumulation of NOS ( Figure 7C). The specificity of anti-iNOS against O. tauri NOS was confirmed by immunoprecipitating O. tauri NOS with anti-OtNOS and then testing for the presence of O. tauri NOS with anti-iNOS (see Supplemental Figure 3 online).
The exposure of photosynthetic organisms to photoinhibitory intensities of light irradiance results in oxidative stress due to the imbalance between the light energy absorbed and the maximum energy that can be used in photosynthesis. NO production in O. tauri challenged by a 103 increase in irradiance (from 40 to 400 mmol m 22 s 21 ) was investigated. It has been reported that this large and sudden increase in light irradiance provokes reversible photoinhibition (Six et al., 2009). Figure 7D shows that NO production was induced ;1.75-fold in photoinhibited O. tauri cell cultures challenged by 400 mmol m 22 s 21 compared with the control cells maintained at 40 mmol m 22 s 21 . NO production returned to almost basal levels when the stressed culture was irradiated at 40 mmol m 22 s 21 for 18 h ( Figure 7D).

NOS Is Present in Photosynthetic Organisms
In this work, we characterized a NOS from O. tauri, a green alga of class Prasinophyceae that branches near the base in the phylogenetic tree of photosynthetic organisms. The O. tauri NOS sequence contains the characteristic oxygenase and reductase domains joined by a CaM binding domain. O. lucimarinus, another species of the Ostreococcus genus, also contains a NOS gene, the expression of which was validated by EST analysis (Lanier et al., 2008). Whereas the O. lucimarinus NOS sequence has just one intron (Lanier et al., 2008), that of the predicted O. tauri NOS sequence (Derelle et al., 2006) contains a second micro-intron (41 nucleotides). It was not verified if this microintron is present in the O. tauri NOS gene or if it is predicted as a consequence of an error in the sequencing process. Thus, the recombinant O. tauri NOS sequence used for biochemical studies in this work may not be the same as the native O. tauri NOS sequence. The predicted micro-intron in O. tauri NOS sequence is located between the binding regions of FAD pyrophosphate and FAD isoalloxazine. Since it does not involve a domain that is critical for cofactor binding, it appears not to affect the activity of the recombinant O. tauri NOS protein.

Phylogenetic Analysis Reveals That O. tauri NOS Clusters with NOS from Cyanobacteria and Amoebozoa as an Outgroup of Metazoa NOS
The study of the phylogenetic relationship between O. tauri NOS and known NOS sequences suggests that O. tauri NOS is an  (Martasek et al., 1996). a Activity assay in presence of all cofactors.
evolutionarily close relative of NOS from animals. O. tauri NOS appears in a well-supported cluster with Synechococcus PCC 7335 and P. polycephalum, a representative of Cyanobacteria and Amoebozoa, respectively. One interesting finding is that the NOS gene is only present in one out of the 13 Synechococcus genomes that have been completely sequenced. Evidence suggests that, in marine picocyanobacteria, the full complement of genes exists in a supragenome, since no single isolate contains the full complement of genes (Scanlan et al., 2009). The high levels of diversity of the gene complement and the efficiency of horizontal gene transfer (HGT) have been postulated to play major roles in the metagenome and population diversity of coastal Synechococcus (Palenik et al., 2009). Synechococcus PCC 7335 appears to be an unusual unicellular organism, since it is the only nonheterocystous N 2 -fixing cyanobacterium currently assigned to this genus (Bergman et al., 1997). It will be interesting to study the role of NOS-dependent NO production in this strain.
A homolog of the Synechococcus PCC 7335 NOS gene is found in the fresh water bacterium Spirosoma linguale but not in other related organisms. Similarly, a homolog of P. polycephalum NOS (Messner et al., 2009) is not found in the completely sequenced genomes of closely related species, such as Dictyostelium discoideum. These findings could suggest that the NOS gene may be mobile within different unicellular organisms as a result of HGT. HGT is considered to be an important evolutionary force that modulates eukaryotic genomes (Keeling and Palmer, 2008). In Ostreococcus, two chromosomes (2 and 19) are different from the remaining 18 in terms of organization and C+G content. It has been suggested that the entire chromosome 19 may have been derived from an exogenous source by HGT (Derelle et al., 2006). Nevertheless, gene loss events cannot be discarded as a possible explanation for the available data.
Different structures of NOS genes are present in terrestrial and aquatic environments. However, there is currently not enough information to establish a correlation between habitats and the presence and structure of NOS genes in different organisms. It will be interesting to investigate the evolutionary forces that link ecological niches with the molecular evolution of NOS.

Recombinant O. tauri NOS Is an Active and Functional Enzyme
The recombinant O. tauri NOS produced in E. coli was a soluble and functional enzyme that exhibited a high level of enzyme activity, similar to that of macrophage iNOS from mouse (Table 1; Stuehr et al., 1991). The K m value of recombinant O. tauri NOS for L-Arg is 12 6 5 mM, which is within the range of previously characterized NOS enzymes (1 to 22 mM; Bredt and Snyder, 1990;Stuehr et al., 1991;Roman et al., 1995;Gerber et al., 1997). We cannot state whether the enzyme is catalytically active as a monomer or dimer. A few studies indicate that the constitutive NOS in animals can be active as a monomer (Bredt and Snyder, 1990;Mayer et al., 1991), while macrophage iNOS is only active as a dimer (Stuehr et al., 1991). The presence of the four regions involved in the dimerization interface (Bird et al., 2002) supports the hypothesis that the active form of O. tauri NOS is a dimer. A dimer of the O. tauri NOS oxygenase domain was modeled and resembles that of human iNOS (see Supplemental Figure 4 online). The C-(x) 4 -C zinc binding motif is strictly conserved across NOS sequences and is required for dimer stability (Hemmens et al., 2000). The zinc atom is tetrahedrically coordinated to two C residues from each NOS subunit. A C-(x) 3 -C motif has thus far only been described in NOS proteins from O. tauri and O. lucimarinus. However, the zinc finger domain present in the GATA family of transcription factors is C-(x) 2 -C (Aita et al., 2000), and the zinc finger domain of the cytochrome oxidase has a highly conserved C-(x) 3 -C motif (Jaksch et al., 2001), indicating that the C-(x) 3 -C motif present in O. tauri NOS could bind zinc. Moreover, the amino acids P and R are present within (x) 3 , between the two C residues in the putative zinc binding domains [C-(x) 3 -C] of NOS from Ostreococcus species and are conserved among most eNOS proteins described to date. Nevertheless, both the dimerization of O. tauri NOS subunits and the binding of zinc remain to be experimentally verified.
Macrophage iNOS binds CaM even at very low Ca 2+ concentrations; thus, iNOS activity is Ca 2+ -CaM independent (Mayer and Hemmens, 1997;Aoyagi et al., 2003). The results obtained from activity measurements and inhibitory assays in this work suggest that O. tauri NOS activity is mostly Ca 2+ -CaM independent. The ACE segment that blocks electron flow in the absence of CaM is not present in O. tauri NOS or in any iNOS described but is present in nNOS and eNOS (Salerno et al., 1997).
The expression of recombinant O. tauri NOS in E. coli generates NO in vivo. It was demonstrated that the NO produced by bacterial NOS is required for maintaining a normal rate of cell growth at the beginning of the stationary phase (Gusarov et al., 2008). The addition of H 2 O 2 induced an increase in NO production in E. coli transformed with the empty vector; however, this basal NO increase failed to protect the cells against the harmful effects of H 2 O 2 . Since E. coli lacks NOS, the evolution of NO following H 2 O 2 treatment could be due to the ability of certain nitrite reductases to reduce nitrite to NO. NO production by E. coli cultures was reported elsewhere (Hutchings et al., 2000;Corker and Poole, 2003). Moreover, H 2 O 2 is able to generate NO nonenzymatically from L-or D-Arg (Nagase et al., 1997;Gotte et al., 2002). E. coli has a complex set of responses to H 2 O 2 and O 22 that involves ;80 inducible proteins and is regulated by the redox-sensitive transcriptional regulator SoxR (Greenberg et al., 1990;Liochev et al., 1994). Interestingly, the E. coli SoxRregulated redox stress response was found to be induced by NO and to confer bacterial resistance to activated murine macrophages (Nunoshiba et al., 1993).

NOS Is Expressed in O. tauri and Is Functionally Influenced by Light Irradiance and Growth Phase of the Microalga
O. tauri cells have a high level of NOS-dependent NO production during the exponential growth phase of the culture. This NO production can be augmented by the exogenous addition of L-Arg. Several studies demonstrate that L-Arg availability is an important factor in NO production (El-Gayar et al., 2003;Lee et al., 2003). Endogenous L-Arg availability may be regulated by (1) increasing the de novo synthesis of L-Arg, (2) increasing L-Arg transport across the cell membrane, and (3) reducing L-Arg breakdown by arginase (Hallemeesch et al., 2002;Flores Figure 7. NO Generation Is Light Irradiance and Growth Phase Dependent. (A) O. tauri cells were grown at 40 mmol m À2 s À1 light irradiance. Culture growth was determined by measuring the OD at 660 nm. NO production was determined using DAF-FM DA with or without the addition of 5 mM L-Arg. Asterisks indicate a statistically significant difference (t test, P < 0.05). Error bars denote SE (n = 3). AU, arbitrary units; DW, dry weight. (B) O. tauri cells grown at 40 mmol m À2 s À1 of light irradiance for 10 d were transferred to 100 mmol m À2 s À1 for 24 h. NO production was determined in the presence or absence of 5 mM L-Arg using DAF-FM DA. Data are expressed as fold increase (arbitrary units per min) with respect to the control. Asterisks indicate a statistically significant difference (t test, P < 0.05). Error bars denote SE (n = 3). (C) Immunoblot analysis of NOS expression in O. tauri cultured for 2, 5, and 10 d at 40 mmol m À2 s À1 light irradiance or 7 d at 40 mmol m À2 s À1 and shift to 100 mmol m À2 s À1 for another 3 d (7 / 10). In another experiment, cells were grown at 100 mmol m À2 s À1 light irradiance for 10 d and treated with or without 5 mM L-Arg for 1 h. Ot NOS protein was detected with a commercial anti-iNOS antibody. The ribulose-1,5-bisphosphate carboxylase/ oxygenase large subunit (RbcL) stained with Ponceau was included as a protein loading control. Relative quantification of Ot NOS and RbcL from blots using Image J software is shown in the bar graph. (D) O. tauri cells grown at 40 mmol m À2 s À1 were transferred to 400 mmol m À2 s À1 for 1 h and then returned to 40 mmol m À2 s À1 for 18 h. NO production was assayed using the fluorescent probe DAF-FM DA. Data are expressed as fold increase (arbitrary units per min) with respect to cells grown at 40 mmol m À2 s À1 . Different letters indicate a statistically significant difference (ANOVA, P < 0.05). Error bars denote SE (n = 3). et al., 2008). Several reports confirm that microalgae excrete free amino acids into the media (Martin-Jé zé quel et al., 1988;Penteado et al., 2009). Our results demonstrate that the increase in external L-Arg concentration can significantly enhance NOSdependent NO production in Ostreococcus. This may be related to the development and ecology of algal blooms in picophytoplankton (Mayali and Azam, 2004;Tillmann, 2004).
NO can participate at different levels during algal bloom successions. Expression of the death-specific protein in the marine diatom Skeletonema costatum is induced by NO, high irradiance, and photoinhibitory compounds (Chung et al., 2008). Vardi et al. (2006) have shown that, in marine diatoms, reactive aldehydes promote the production of large amounts of NO through a NOS-like activity and induce cell death. In a more recent work, Vardi et al. (2008) also demonstrated that a high level of NO production associated with a GTP binding protein was responsible for reduced growth and impaired photosynthetic efficiency in the marine diatom Phaeodactylum tricornutum. Thus, several lines of evidence indicate that augmented NO generation is closely related to the regulation of algal bloom evolution. Hence, L-Arg-dependent NO production during the life cycle of Ostreococcus points to the potential impact of amino acid excretion on both the algal bloom and the marine nitrogen (N) cycle. N is the only essential element whose concentration in seawater is controlled by biological activity (Morel, 2008).
Marine photosynthetic organisms confront environmental changes, such as changes in light and temperature, during the day. Irradiation of a photosynthetic organism with photoinhibitory light intensities provokes the photoinactivation of photosystem II (PSII) and generates an oxidative stress response that can damage biomolecules. To counter photoinactivation, O. tauri removes the photoinactivated D1 protein from PSII and replaces it through de novo synthesis and reassembly of PSII (Six et al., 2009). NO is a potent antioxidant molecule that was shown to protect lipids, proteins (including D1 protein), and nucleic acids from the photooxidative damage generated by bipyridinium herbicides in higher plants (Beligni and Lamattina, 2002). The induction of NOS activity in O. tauri upon exposure to photoinhibitory light intensities may reduce oxidative damage and promote the repair of this damage.
This report provides compelling evidence that an active NOS functions in a photosynthetic organism belonging to the plant kingdom and that this enzyme contains the main characteristics of the NOS enzymes present in animals.

Materials
L-Arg, 295-ADP-agarose, CaM, H 4 B, TFP, hemoglobin, anti-iNOS antibody from mouse, and recombinant mouse iNOS (16.8 units/mg protein) were purchased from Sigma-Aldrich. The NOS inhibitors L-NAME, L-NMMA, and L-NNA were obtained from Calbiochem. The DNA clone for the coding sequence of the NOS gene from Ostreococcus tauri (Ot NOS) and the polyclonal anti-OtNOS antibody were acquired from Genscript. Restriction enzymes, Bacto-yeast extract, IPTG, and Escherichia coli DH5a cells were purchased from Life Technologies. BL21 (DE3) pLys cells and pET24b were from Novagen. O. tauri (strain OTTH 0595) was purchased from the Roscoff Culture Collection (Station Biologique, Roscoff, France).

Homology Modeling
Fold assignment was performed using FFAS03 (Jaroszewski et al., 2005) and HHPred (Sö ding et al., 2005). Structural models were built with the program MODELER v 9.5 (Sali and Blundell, 1993) using ClustalX-and FFAS03-derived alignments with explicit consideration for the presence of ligands and the oligomeric state. The template was selected using FFAS03 and corresponds to the crystal structure of the NOSoxy region of Bos taurus eNOS and of the NOSred region of nNOS from Rattus norvegicus. The model was validated using PROSA II software (Wiederstein and Sippl, 2007), and the figures were drawn using Web Lab ViewerLite 3.20 software (Molecular Simulations). Structural superposition of O. tauri NOS and B. taurus eNOS was performed with the program SuperPose (Maiti et al., 2004). O. tauri NOS and human NOS were aligned using ClustalX software (version 1.81; Thompson et al., 1997) and edited with GeneDoc software (version 2.5.010).

Sequence Data and Phylogenetic Analysis
Protein sequences with similarity to O. tauri NOS (E-value < 1.10 2110 ) were retrieved using BLAST and the nonredundant database from GenBank (see Supplemental Table 1 online). CDHIT software (Huang et al., 2010) was used to remove all sequences sharing >90% identity. The sequences were aligned with ClustalX and edited with GeneDoc software (see Supplemental Data Set 1 online). Domain organization was analyzed using InterPro database (Hunter et al., 2009) and PFAM (Finn et al., 2008). Maximum likelihood phylogenetic analyses were performed with PHYML (Guindon and Gascuel, 2003). The best evolutionary model for the phylogenetic inference was estimated using ModelTest (Posada and Crandall, 1998). Nonparametric bootstrapping (1000 replicates) was used to assess tree branching support. The program iTool (Letunic and Bork, 2007) was used to display phylogenetic trees (http://itol.embl.de/).

DNA Manipulation
The O. tauri NOS coding DNA sequence was synthesized, sequenced, and cloned into pUC57 using XbaI and XhoI enzymes. pUC57-OtNOS was digested with BamHI and SacI to yield a fragment of 3258 bp. The product was purified and subcloned into pET24b. The ligation product was used to transform E. coli DH5a. Colonies were screened by PCR using the following primers: forward, 59-GCTGGGCGCCGGAAAA-GAC-39, and reverse, 59-GCGCCGGCCGAAACTCAAC-39. The expected product length was 2031 bp. pET24b-OtNOS was used to transform BL21 protease-deficient E. coli via electroporation. This strain is deficient in the proteases ompT and lon, resulting in a suitable host for optimal recombinant expression (Sørensen and Mortensen, 2005).

Protein Expression and Purification
Fernbach flasks containing 1 liter of modified Terrific Broth (20 g of yeast extract, 10 g of bactotryptone, 2.65 g of KH 2 PO 4 , 4.33 g of Na 2 HPO 4 , and 4 mL of glycerol) and kanamycin (50 mg/mL) were inoculated with 1 mL of culture and shaken at 190 rpm at 308C. Recombinant protein expression was induced at OD 600 = 0.6 by the addition of 0.5 mM IPTG. The heme and flavin precursors 8-aminolevulinic acid and riboflavin were added to final concentrations of 450 and 3 mM, respectively.
Cells were harvested after 40 h of induction and were resuspended in 30 mL of buffer (100 mM Tris-HCl, pH 7.4, 1 mM EDTA, 1 mM DTT, 10% [v/v] glycerol, 1 mM PMSF, 5 mg/mL leupeptin, and 5 mg/mL pepstatin) per liter of initial culture and lysed by pulsed sonication (six cycles of 20 s). Cell debris were removed by centrifugation, and the supernatant was applied to a 29,59-ADP-agarose 4B column (1 mL) equilibrated in buffer B (50 mM Tris-HCl, pH 7.4, 0.1 mM EDTA, 0.1 mM DTT, 10% glycerol, and 100 mM NaCl). The column was extensively washed with 10 column volumes of buffer B and finally with buffer B and 500 mM NaCl. The protein was eluted with buffer B, 500 mM NaCl, and 25 mM 2'-AMP.

Determination of NOS Activity
NO synthesis was assayed using two methods: (1) oxyhemoglobin and (2) [ 3 H]L-citrulline formation. The oxyhemoglobin method was performed as described by Ghafourifar et al. (2005). Hemoglobin was completely reduced to oxyhemoglobin with sodium dithionite. The concentration of oxyhemoglobin in the solution was determined using a molar extinction coefficient of 131 mM 21 cm 21 at 415 nm. A 500-mL reaction containing 20 mM oxyhemoglobin, 7.5 mM HEPES-NaOH, pH 7.5, 5 mM DTT, 100 mM L-Arg, 1 mM NADPH, 10 mM CaCl 2 , 10 mM CaM, 100 mM H 4 B, and 100 units/mL catalase was prepared. The reaction was initiated by the addition of 0.5 mM purified O. tauri NOS protein. The NO-dependent conversion of oxyhemoglobin to methemoglobin was monitored on a spectrophotometer (Ultrospec 1100 pro; Amersham Biosciences) by scanning between 380 and 450 nm. An extinction coefficient of 100 mM 21 cm 21 between the peak at 401 nm and the valley at 420 nm was used to quantify NO production. Citrulline formation was determined as previously described (Bredt and Snyder, 1990). Enzymatic reactions were conducted at 258C in 50 mM Tris-HCl, pH 7.4, containing 50 mM L-Arg, 1 mCi [ 3 H]Arg monohydrochloride (40 to 70 Ci/mmol; Perkin-Elmer), 100 mM NADPH, 10 mM FAD, 2 mM CaCl 2 , 1 mg CaM, and 100 mM H 4 B in a volume of 40 mL. Enzymatic reactions were initiated by adding 0.5 mM NOS and terminated after 30 min by the addition of 400 mL of ice-cold 20 mM sodium acetate, pH 5.5, containing 1 mM L-citrulline, 2 mM EDTA, and 0.2 mM EGTA (stop buffer). Samples were applied to columns containing 1 mL of Dowex AG50W-X8, Na + form (Bio-Rad; 100 to 200 mesh), preequilibrated with stop buffer. L-citrulline was eluted with 2 mL of distilled water. Aliquots of 0.5 mL of eluate were dissolved in 10 mL of scintillation liquid, and radioactivity was measured in a Beckman LS 3801 liquid scintillation system. The formation of L-citrulline was verified by thin layer chromatography.
NADPH oxidation was measured in a 500-mL volume containing 0.5 mM O. tauri NOS, 50 mM Tris-HCl, pH 7.6, 5 mM DTT, 100 mM L-Arg, 1 mM NADPH, 10 mM CaCl 2 , 10 mM CaM, 10 mM H 4 B, and 100 units/mL catalase. The rate of decrease in absorbance at 340 nm was monitored for 10 min at 258C using a spectrophotometer. An extinction coefficient of 6.22 mM 21 cm 21 at 340 nm was used to calculate NADPH oxidation.

NO Production in Cultures
O. tauri cultures were grown in Erlenmeyer flasks containing K medium (Keller et al., 1987) at 20 6 18C under a 12 h:12 h (light:dark) photoperiod. Light experiments were conducted over irradiances ranging from 40 to 400 mmol m 22 s 21 . NO content in the E. coli or O. tauri cultures was quantified using the NO-sensitive fluorescence probe DAF-FM DA (Invitrogen). DAF-FM DA (10 mM) was added to the culture medium 20 min before measurement. NO production was initiated by the addition of the substrate L-Arg. NO fluorescence intensity (excitation 495 nm; emission 515 nm) was measured using a fluorescence plate reader (Fluoroskan Ascent; Thermo Electron) or visualized under an inverted fluorescence microscope (Nikon Eclipse TI). NO formation in E. coli was also measured using the Griess reagent for nitrate measurement as described by Xu et al. (2000).

Cell Death Analysis
E. coli cultures were treated with 30 mM H 2 O 2 for 30 min. Cells were stained with 2 mM of the membrane-impermeable dye SYTOX (Molecular Probes) for 10 min in the dark at room temperature. Cells were washed with 1.5 volumes of fresh medium. Fluorescence intensity (excitation 480 nm; emission 525 nm) was measured using a fluorescence plate reader (Fluoroskan Ascent).

Statistical Analysis
Results are expressed as mean 6 SE. Statistical analysis was performed employing SigmaStat statistical software (Jandel Scientific) using analysis of variance (ANOVA) for multiple comparison analyses and the t test for pairwise comparisons.

Accession Numbers
Sequence data from this article can be found in the GenBank/EMBL database under the following accession numbers: O. tauri NOS (CAL57731), eNOS (NP_000594), nNOS (NP_000611), iNOS (AAI30284), and Spirosoma linguale NOS (YP_003391026). The accession numbers for the amino acid sequences used in the phylogenetic analysis can be found in Supplemental Table 1 online. The crystal structures from this article can be found in the Protein Data Bank (www.rcsb.org) under the following codes: R. norvegicus nNOSred (1tll) and B. taurus eNOSoxy (1fop).

Author Contributions
Experiments were designed by N.F., N.C.-A., G.P., G.C., G.S., and L.L. and conducted by N.F., G.C., and N.C.-A. Bioinformatic analyses were conducted by G.P., N.C.-A., N.F., and L.L. The manuscript was prepared and written by N.F., N.C.-A., and L.L. The design, supervision, and direction of the project were performed by L.L.

Supplemental Data
The following materials are available in the online version of this article. Supplemental Table 1. Full Name and GI Number of All NOS Sequences Retrieved with BLAST and the Nonredundant Database from GenBank.
Supplemental Data Set 1. Alignment of NOS Proteins Used to Build the Phylogenetic Tree Presented in Figure 3.