Nitric oxide synthases (NOS, EC 18.104.22.168) are enzymes that catalyze the formation of nitric oxide (NO) from L-arginine. NO is an important signaling molecule that participates in a number of biological processes, including neurotransmission, vasodilation, and immune response (1). NOS proteins are active as homodimers, and each monomer consists of two well-defined domains. The N-terminal oxygenase domain binds heme, L-arginine, and (6R-)5,6,7,8-tetrahydrobiopterin (H4B). The C-terminal reductase domain binds flavin adenine dinucleotide (FAD), flavin mononucleotide (FMN), and nicotinamide adenine dinucleotide phosphate (NADPH) and provides electrons to the oxygenase domain from NADPH. Both domains can be expressed and purified separately, and conserve their ligand binding and reactivity. Between the oxygenase and reductase domains, a short sequence (30–40 aminoacids) provides a binding site for calmodulin (CaM) (Fig. 1).
The oxygenase domain of NOS catalyzes the oxidation of L-arginine to L-citrulline and NO in two sequential oxidation steps (Scheme 1). In the first step, L-arginine is hydroxylated to make Nω-hydroxy-L-arginine (NOHA) in a process that requires one molecule of NADPH and one molecule of oxygen per mol of L-arginine reacted. In the second step, NOHA is oxidized to L-citrulline and NO and 1 molecule of oxygen and 0.5 molecules of NADPH are required.
The three mammalian isoforms of NOS show high sequence homology but greatly differ in their localization and regulation. The inducible isoform (iNOS) has a high affinity for CaM and is constitutively active; this isoform is mainly regulated at the transcriptional level. Endothelial (eNOS) and neuronal (nNOS) enzymes are constitutively expressed and are reversibly activated by calcium through the binding of Ca2+-containing CaM. Related proteins have been discovered in bacteria but consisting of only the oxygenase domain (2). The bacterial NOS enzymes also contain heme but are less stringent in their pterin requirements and can often bind tetrahydrofolate (H4F) and H4B with similar affinity. Due to the absence of a connected reductase domain, they must receive electrons from other electron transfer proteins (2–4). The recently discovered NOS from Sorangium cellulosum is a notable exception to this rule (5).
The general biochemistry of NOS enzymes has been extensively reviewed from the biophysical (6–9) and clinical perspective (10). Reviews on H4B biochemistry have also treated NOS (11, 12). In this review, we will focus on the function of tetrahydrobiopterin in NOS enzymes.
H4B Requirement for NOS Catalysis
Early studies of NOS enzymes identified H4B as a necessary cofactor for NO synthesis (13–15) with a stoichiometry of one molecule of H4B per subunit (16). Nevertheless, the function of H4B was controversial, with a variety of suggested roles. A redox cycle involving H4B and H2B, as in aromatic amino acid hydroxylases was considered, but this would require one H4B molecule per cycle, in contradiction with experimental data (15, 17). Strong evidence indicated that H4B enhanced dimer formation (18, 19). However, the evidences of some redox effect did slowly accumulate. Single turnover experiments indicated that the stability of the heme-oxy complex was dependent on H4B presence, and H4B was required for the reaction to advance past the formation of a ferrous dioxygen/ferric superoxide complex (20). Other results indicated that H4B analogs were able to catalyze L-arginine hydroxylation to a variable extent, but all were reduced forms (21). It was even unclear if H4B was required for the NOHA oxidation step (22). The detection of a H4B radical (23–26) and the coupling of this radical formation with product formation (26) finally lead to the currently accepted notion—the decay of the ferrous dioxygen/ferric superoxide complex is linked to the formation of a H4B radical species (8, 9, 27, 28).
H4B Binding in NOS
Pterin Binding Affinity of NOS
NOS enzymes bind H4B with high affinity. The presence of significant amounts of H4B in the enzyme after purification evinces the tight binding of the cofactor by the native protein (13–15). The binding of L-arginine and H4B shows a synergistic effect: prior binding of L-arginine increases the binding affinity for H4B and vice versa (21, 29, 30). Mammalian NOS enzymes bind H4B with affinities in the nM range. H4B is usually the preferred substrate over H2B, except for the eNOS enzyme (31). This fact has important implications for human pathology (10, 12, 31). Bacterial NOS enzymes operate in a wide range of KD values, and sometimes bind H4F with higher affinity (2, 32, 33). It should be noted that although many other pterins can bind to NOS (see (11) and references therein), only a few H4B analogs can support NO synthesis (21). The binding affinities for several pterins are shown in Table 1.
Table 1. Dissociation constants for pterin binding to NOS enzymes
Pterin Binding Pocket in Mammalian and Bacterial NOS
Structures of the oxygenase domains for the three eukaryotic NOS isoforms (34–36) and several bacterial NOS proteins (37–39) have been reported. The alignment of the sequences for these NOS proteins indicates the notable similarity between mammalian and bacterial NOS enzymes (Fig. 2). The presence of additional elements in the N-termini of the mammalian proteins can be also noted; these motifs are involved in dimer formation and pterin binding (Figs. 2 and 3). The available structures clearly show that the absence of this N-termini allows bacterial NOS proteins to bind larger pterins (Fig. 3). The conservation of the H4B binding environments in eNOS, iNOS, and nNOS is remarkable. Moreover, the bacterial NOS proteins also share many similarities in the pterin binding region. H4B is bound to NOS by an extensive network of hydrogen bonds, with a highly conserved pattern for the three NOS isoforms (34–36) (Fig. 3). Most hydrogen bonds are contributed by the protein but there is also an interaction with the heme group. Several hydrogen bonds are contributed by the protein backbone and the side chain of a conserved Arg residue (Arg375 in mouse iNOS). A conserved tryptophan residue also provides stabilization through a π-stacking interaction (Trp457 in mouse iNOS). Notably both Arg and Trp residues are conserved in mammalian and bacterial NOS proteins. The pterin binding pocket is placed in the oxygenase dimer interface. This location clearly suggest a relationship with the role of H4B binding on stabilization of the NOS dimer.
H4B Function in NOS Catalysis
H4B Role in the L-Arginine Hydroxylation Step
Despite initial difficulties to assess the role of H4B in NOS catalysis, the study of the formation of NOHA from L-arginine yielded increasing evidence of a role for H4B beyond the H4B–H2B cycle of aromatic hydroxylases. The presence of H4B increases the rate of the FeIIO2 complex decay, a novel role for H4B (20). Other studies observed that NOHA formation only occurred in the presence of H4B (40, 41). As previously pointed out, the observation of a H4B radical formation during the reaction of NOS with L-arginine clearly indicated the active role of H4B during the formation of NOHA from L-arginine (23–26). The observation of the build-up of an H4B radical as the ferrous oxygen complex decays was instrumental in establishing a coherent mechanism (26) (Fig. 4).
First steps of the reaction were clear, with the reductase domain of NOS providing the electrons to reduce the heme, and subsequent oxygen binding to form a ferrous-dioxygen complex (FeIIO2). As this species decays with concomitant formation of an H4B radical, a one electron transfer from H4B to the heme-oxy complex with the formation of a ferric peroxo species is expected. From this point, different reactions are possible. By analogy with P450 systems, protonation of this species to form the ferric-hydroperoxo species and eventual formation of the FeIVO porphyrin radical (Compound I) species was proposed (Fig. 5). In this process, the H4B radical is still present after L-arginine hydroxylation, and an additional electron—provided by the flavoprotein domain (42)—is needed to return to the initial resting state.
H4B Role in the NOHA Oxidation Step
The formation of citrulline and NO from NOHA has many similarities with the process of L-arginine oxidation to NOHA and also substantial differences. We must consider that the oxidation of L-arginine to NOHA requires two electrons and two protons, but the reaction of NOHA to citrulline and NO requires one electron, as a two electron donation would lead to the formation of nitroxyl (NO−) species as final product. Therefore, it was considered that the electron transfer from H4B may not be necessary, and even detrimental, for this step.
Extensive evidence further indicates that H4B is necessary for NOHA oxidation (43–45). Interestingly, reaction of nNOS with NOHA and hydrogen peroxide in the absence of H4B leads to the formation of nitroxyl, whereas H4B directs the reaction toward NO instead (46). Therefore, it was proposed that the H4B radical formed in the first steps of the reaction could actually be reduced by a FeII–NO intermediate to produce the observed FeIII-NO species (24, 46, 47). Later reports are consistent with such model (44, 48–50).
These observations indicate that the requirement for H4B in the NOHA oxidation step appears to be two-pronged (Fig. 6). First, H4B it is needed to transfer an electron to the ferrous-dioxygen complex, as for the L-arginine hydroxylation step. Then, after the reaction of the ferric hydroperoxo species with NOHA, a ferrous nitroxyl complex is formed. The H4B radical will be reduced by this species giving the product ferric–NO complex and regenerating the H4B.
Effect of Mutations in the H4B Binding Pocket
The effect of changes in the H4B binding pocket illustrates the fine tuning of the H4B reactivity in NOS. The mutation of iNOS Trp457 by Phe or Ala decreases the rate of H4B radical formation and accelerates its decay. The ferrous-dioxygen complex also decays at a slower rate similar to that of H4B radical formation, as expected from the mechanism. The amount of H4B radical stabilized by Trp457Phe and Trp457Ala is around 75% of wild-type enzyme. These changes cause a decreased NOHA yield (51). Similar effects are observed for the nNOS mutants Trp678Phe and Trp678Ala, indicating a conserved role for this residue in the different isoforms (52). NOHA single turnover experiments have been also reported for the iNOS mutants Trp457Phe and Trp457Ala. The mutants show a decrease in the reaction rate, no detectable formation of H4B radical or FeIIINO species, and decreased citrulline yield (53). The mutants are able to produce NO, to a lesser extent than the wild-type enzyme. These observations suggest that the lack of FeIII–NO species is due to a rate of formation slower than that of NO dissociation from the FeIII–NO complex, unlike the wild type enzyme where enough FeIII–NO accumulates to be detected. The mutations cause the FeII–O2 complex to decay at a slower rate; this decrease correlates with a decreased citrulline yield. These results confirm that modification of the H4B radical stability impacts the reaction of H4B with the FeII–O2 complex in both reaction steps.
Replacement of iNOS Arg375 by Lys, Asn, or Asp also causes a decrease in the reaction yield. The underlying causes appear more complex in this case. For the Arg375Asn and Arg375Asp, these effects are partly related to impaired formation of the NOS dimer and diminished ability to stabilize the H4B radical. Alternatively, the Arg375Lys mutant is able to form the H4B radical at a faster rate than wild-type iNOS. However, Arg375Lys showed lower NOHA yield, decreased NO synthesis, and increased NADPH oxidation. Thus, fast H4B reactions can lead to increased uncoupling of NADPH oxidation and NO synthesis. These results stress the need for a timely electron transfer for NO synthesis (54).
Our knowledge about H4B function in NOS has rapidly evolved. In less than 25 years, H4B has gone from an exotic cofactor of unknown function to a specialized electron donor with a role never seen before in H4B dependent enzymes. Some questions about H4B function still remain, the possible role of H4B as proton donor during catalysis or the exact protonation role of the H4B species is still discussed (8, 9, 28). The details of the reaction in bacterial NOS, although largely similar to mammalian NOS, can yield some surprises (2, 45).