Mammalian NOSs (nitric oxide synthases) are haem-based monoxygenases that oxidize the amino acid arginine to the intracellular signal and protective cytotoxin nitric oxide (NO). Certain strains of mostly Gram-positive bacteria contain homologues of the mammalian NOS catalytic domain that can act as NOSs when suitable reductants are supplied. Crystallographic analyses of bacterial NOSs, with substrates and haem-ligands, have disclosed important features of assembly and active-centre chemistry, both general to the NOS family and specific to the bacterial proteins. The slow reaction profiles and especially stable haem-oxygen species of NOSs derived from bacterial thermophiles have facilitated the study of NOS reaction intermediates. Functionally, bacterial NOSs are distinct from their mammalian counterparts. In certain strains of Streptomyces, they participate in the biosynthetic nitration of plant toxins. In the radiation-resistant bacterium Deinococcus radiodurans, NOSs are also likely to be involved in biosynthetic nitration reactions, but, furthermore, appear to play an important role in the recovery from damage induced by UV radiation.
- bacterial pathogenesis
- nitric oxide synthase (NOS)
- radiation resistance
Mammalian NOS (nitric oxide synthase)
In mammals, nitric oxide (NO) is involved in many biological processes that range from protection against pathogens to blood pressure regulation and nerve cell transmission . NO is produced by three isoenzymes of the NOSs that all catalyse the oxidation of L-arginine to NO (Figure 1) . Mammalian NOSs are homodimers that contain an N-terminal haem oxygenase domain (NOSox) and a C-terminal flavoprotein reductase domain NOSred. NOSox binds L-arginine, haem and the redox-active cofactor H4B (6R-tetrahydrobiopterin) and comprises the catalytic centre of the enzyme, whereas NOSred binds FAD, FMN and NADPH and functions to supply reducing equivalents for oxygen activation. A CaM (calmodulin)-binding sequence links NOSox and NOSred and regulates reduction of NOSox by NOSred in a Ca2+-dependent manner . The catalytic mechanism of NO formation by mammalian NOS (Figure 1) involves two haem-based oxidations: conversion of L-arginine into the stable intermediate NOHA (Nω-hydroxy-L-arginine) and then conversion of NOHA into NO and L-citrulline . The first reaction has largely been assumed to proceed analogously to the mono-oxygenase reactions of the cytochromes P450, whereas a number of mechanisms have been proposed for the second reaction; in truth, many key issues remain to be resolved for both steps . To distinguish further the NOSs from the cytochromes P450, both catalytic steps require H4B (Figure 2), which rapidly supplies electrons to the ferrous-oxy complex during oxygen activation [3,5]. Although rapid reduction of the ferrous-oxy species is necessary to out compete autoxidation, electron transfers from NOSred must be slow to prevent reduction of product Fe(III)–NO complex to Fe(II)–NO, whose stability disfavours NO release from the haem centre . Thus both fast and slow electron transfers are needed during catalysis, and this requirement has been met by incorporating pterin as a reversible electron donor.
Genome sequencing efforts first identified bacterial protein sequences with substantial similarity to those of the mammalian NOSs. A number of groups have now characterized the spectroscopic properties, catalytic profiles and structures of bacterial NOS proteins derived from largely Gram-positive bacteria [7–19]. Bacterial NOSs found in strains of Bacillus, Deinococcus, Staphylococcus and Streptomyces strains are homologous with the mammalian NOSox, but lack an associated NOSred and an N-terminal β-hairpin hook that binds Zn2+, the dihydroxypropyl side chain of H4B, and the adjacent subunit of the oxygenase dimer [7,9]. Nevertheless, Deinococcus radiodurans NOS and Bacillus subtilis NOS are dimeric, have a normal haem environment, bind the substrate L-arginine and produce NO species in a manner dependent on pterin [either with H4B or with the related cofactor THF (tetrahydrofolate)] [7,8]. The crystal structure of B. subtilis NOS complexed with L-arginine and THF  (Figure 3), and those of Staphylococcus aureus  and Geobacillus stearothermophilus , confirmed that bacterial NOS proteins are structurally similar to mammalian NOSox, with the exception of an absent N-terminal β-hairpin hook and the Zn2+-binding region  (Figure 3). Conservation of key residues among bacterial and mammalian NOSs also suggested that all bacterial NOSs produce NO species from L-arginine and NOHA . Bona fide NO production has been demonstrated for B. subtilis NOS [7,8] and G. stearothermophilus NOS . Interestingly, the lack of the N-terminal hook allows bacterial NOSs to bind larger pterin cofactors than the mammalian NOSs (i.e. THF). This may be related to the fact that some NOS-containing bacteria such as D. radiodurans do not appear to contain all of the biosynthetic enzymes necessary to produce H4B .
Structure and mechanism
The chemical mechanism of NO production by bacterial NOS relates closely to that of mammalian NOS (Figure 2), with the exception that electrons cannot be supplied from a covalently attached reductase domain. Indeed, a B. subtilis flavodoxin can act as an efficient electron source for B. subtilis NOS . However, genetic experiments have shown that this particular flavodoxin is not necessary for NO production by B. subtilis NOS in cells, and that there are likely to be a number of different reductase proteins that can functionally reduce B. subtilis NOS and support in vivo NO synthesis .
Bacterial NOSs have served as useful subjects for studies into the mechanism of NO production by this general class of enzyme. Nitrosyl–haem complexes of B. subtilis NOS bound to both L-arginine and NOHA have been used to mimic the reactive ferrous-haem oxy species that precede substrate oxidation . These structures suggest that L-arginine facilitates hydrogen-bonding interactions to the terminal oxygen of the ferrous-oxy complex, whereas NOHA, which is protonated on the oxime nitrogen when bound to the enzyme, directs hydrogen bonds to the haem-proximal oxygen. The implication is that, on further haem-oxy reduction, NOHA may stabilize a ferric-peroxo intermediate, whereas L-arginine may facilitate breakdown to a ferryl-oxo species, analogous to peroxidase compound I (Figure 2). Resonance Raman studies of Staphylocccus and Bacillus NOS haem-oxy complexes are consistent with this model [12,13]. One of the most notable differences between mammalian NOS and B. subtilis NOS is that product NO release rates from the haem are considerably lower in B. subtilis NOS . This kinetic difference correlates with a structural difference in the active centre, where a mammalian NOS valine residue that projects over the oxygen co-ordination site is replaced by a larger isoleucine residue in most bacterial enzymes . Indeed, switching of this residue in bacterial and mammalian NOSs reverses the trend in the haem ligand dissociation rate [23,24]. The NOS derived from G. stearothermophilus, a thermophile, has especially stable haem-oxygen intermediates at room temperature (Figure 4) . At cryogenic temperatures, EPR/ENDOR (electron nuclear double resonance) techniques have revealed that the initial G. stearothermophilus NOS oxy-haem intermediates may indeed be different for the two steps of the reaction cycle (R. Davydov, J. Sudhamsu, B.R. Crane and B.R. Hoffman, unpublished work). On temperature annealing, novel paramagnetic species form in the haem centre before products. Similar species were not observed with mammalian NOS, presumably because of their instability . This on-going work holds great promise for revealing new insights into the mechanism of substrate oxidation in both steps of NOS catalysis.
The bacterial NOSs have also aided in the characterization of structural states common to mammalian NOSox that are involved in enzyme assembly and possibly regulation. Mammalian NOSox is known to form a ‘loose dimer’ in the absence of H4B and substrate . The loose dimer has a more exposed haem group, is more susceptible to proteolysis and has a lower haem redox potential compared with the pterin- or substrate-bound enzyme . B. subtilis NOS was crystallized in this form, and its structure revealed that the dimer interface and cofactor-binding site was highly disordered . These structural perturbations propagate to bracketing helical regions and the active centre, where they influence substrate binding. Clearly, cofactor and substrate recognition is tightly coupled with formation of a tight catalytically competent NOS dimer. The loose dimer state may be important for limiting haem-based oxygen reduction in the absence of substrate and cofactor.
Biological functions of bacterial NOS
Bacteria have long been known to be sources of NO. In fact, NO was first identified as a biological product in 1967, when it was shown to be an intermediate in anaerobic denitrification . Moreover, NO from host or environmental sources is an important signal for many bacteria, and, as such, many different sensing systems have been identified [29,30]. However, only a subset of bacteria that respond to NO contain a NOS, and, until recently, the biological function of these enzymes was unknown. The first hint into bacterial NOS function came from the finding that a NOS open reading frame was contained on a pathogenicity island common to certain Streptomyces strains that cause potato scab disease . The pathogenic Streptomyces strains produce a family of plant toxins called thaxtomins that interfere with plant cell wall synthesis. Thaxtomins are unusual dipeptides (derivatives of cyclo-[L-tryptophanyl-L-phenylalanyl]) (Figure 5) produced by non-ribosomal peptide synthesis . Most interestingly, the tryptophan moiety of the phytotoxin is nitrated at the 4-position , and this nitration is carried out, at least in part, by the NOS . Involvement in biosynthetic nitration is an unprecedented metabolic role for a NOS protein. Biosynthetic nitration reactions are rare and usually involve the oxidation of an amine . The chemical mechanism of a NOS-mediated nitration may be complex because NO is unlikely to react directly with indole. Nevertheless, readily oxidized forms of NO, such as nitrosonium (NO+), peroxynitrite (ONOO−), nitronium (NO2+) or nitrogen dioxide (NO2), actively nitrate aromatic amino acids [34,35]. Interestingly, the Streptomyces strains produce NO far in excess of that needed for toxin biosynthesis, and NO production occurs at sites of infection, such as nascent root tips . In the plant, NO is known to act as a signal for the growth and extension of such structures , raising the interesting possibility that the NO produced by the pathogen also promotes the growth of tissue conducive to colonization.
NO and tryptophanyl-tRNA in D. radiodurans
Mechanistic studies of thaxtomin nitration have lagged because of difficulties associated with producing the Streptomyces enzymes in sufficient amounts. However, the D. radiodurans NOS has been shown to catalyse the production of small amounts of 4-nitrotryptophan in vitro, when reductants are supplied by a surrogate NOSred . Furthermore, the D. radiodurans NOS co-purifies with an unusual TrpRS (tryptophanyl-tRNA synthethase) , providing a link to tryptophan metabolism in Deinococcus as well as Streptomyces. D. radiodurans has two TrpRS genes, and, of the two, one (TrpRS I) is more closely related by sequence to standard bacterial TrpRSs than the other (TrpRS II) . Both TrpRS I and TrpRS II amino-acylate tRNATrp, but TrpRS I is ∼3× more active than TrpRS II. Either gene can be deleted and D. radiodurans is viable, but ablation of both is lethal (B. Patel, J. Widom and B.R. Crane, unpublished work). TrpRS II forms a complex with D. radiodurans NOS that increases its nitration activity towards tryptophan . The structure of TrpRS II revealed an unusual tryptophan-binding site that suggested that analogues of tryptophan could be recognized and potentially coupled to tRNA (Figure 6) [40,41]. Indeed, with nearly equal specificity, TrpRS II will charge tRNATrp with tryptophan, 4-nitrotryptophan or 5-hydroxytryptophan . Thus the D. radiodurans NOS–TrpRS II complex can produce 4-nitrotryptophan-tRNATrp. The purpose of this unusual product is currently unknown. Our studies have not found evidence for the incorporation of 4-nitrotryptophan into proteins. Instead, it is more likely that 4-nitrotryptophan-tRNATrp is used for the production of an as yet to be discovered secondary metabolite. Analysis of charged and uncharged tRNATrp pools in D. radiodurans indicates the presence of multiple tRNATrp species (B. Patel, J. Widom and B.R. Crane, unpublished work). Furthermore, these pools are affected by gene deletions for nos or trprsII. Current work is directed at identifying modified tRNATrp in D. radiodurans.
NOS-derived NO in stress responses
Gene deletions of nos have also been studied in B. subtilis  and the human pathogen Bacillus anthracis . The B. subtilis knockouts were shown to be more susceptible to oxidative damage under conditions where reduced thiols were up-regulated . This lead to the hypothesis that NOS-derived-NO reduces oxidative damage by blocking reduced thiols that would normally catalyse Fenton chemistry and thereby produce hydroxyl radicals . Consistent with this, the resistance of B. anthracis to the oxidative burst of macrophages decreases substantially in a nos mutant . Furthermore, NO production is induced in the pathogen when it is attacked by the immune cells, indicating that NO is a pathogen response to host defence . We have also shown that NOS-derived NO plays an important role in the recovery of D. radiodurans from UV radiation and DNA-damaging agents (B. Patel, J. Widom and B.R. Crane, unpublished work). The deficiency in radiation resistance that results from nos gene deletion can be rescued by chemical NO donor compounds. Both NO production and nos gene expression appear to be induced by the damage agent, in this case radiation. How NO is enabling these stress responses in Bacillus spp. and Deinococcus is an open question for which we currently have little data. One compelling hypothesis is that it acts as signal to co-ordinate protection and repair. This leads to the issue of how the nos gene itself is regulated by radiation, redox or host responses.
Studies of bacterial NOSs promise to generate new understanding of the widely employed metabolite NO. These enzymes continue to serve as tractable subjects for unveiling the mechanistic details of NO synthesis. Furthermore, bacterial NOSs appear to be involved in many novel and interesting biochemical pathways. Their role as NO generators may be employed in multiple cellular processes, even within the same organism (e.g. D. radiodurans). In this sense, the global functions of bacterial NOSs parallel their mammalian counterparts; however, it is within the specifics of these biological activities where we can hope to discover novel aspects of NO chemistry and biology.
Transition Metals in Biochemistry: A joint Biochemical Society meeting with the Inorganic Biochemistry Discussion Group to honour Professor Andrew Thomson FRS, held at University of East Anglia, Norwich, U.K., 24–26 June 2008. Organized and Edited by Steve Chapman (Edinburgh, U.K.), David Richardson (University of East Anglia, U.K.) and Nick Watmough (University of East Anglia, U.K.).
Abbreviations: H4B, 6R-tetrahydrobiopterin; NOHA, Nω-hydroxy-L-arginine; NOS, nitric oxide synthase; NOSox, N-terminal haem oxygenase domain; NOSred, C-terminal flavoprotein reductase domain; THF, tetrahydrofolate; TrpRS, tryptophanyl-tRNA synthethase
- © The Authors Journal compilation © 2008 Biochemical Society