HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via CrossRef Citing Articles via Google Scholar Google Scholar Articles by Cabello, P. Articles by Moreno-Vivián, C. Search for Related Content PubMed PubMed Citation Articles by Cabello, P. Articles by Moreno-Vivián, C. Agricola Articles by Cabello, P. Articles by Moreno-Vivián, C.

Nitrate reduction and the nitrogen cycle in archaea

¹ Departamento de Biología Vegetal, Área de Fisiología Vegetal, Universidad de Córdoba, Spain
² Departamento de Bioquímica y Biología Molecular, Edificio Severo Ochoa, 1^a planta, Campus Universitario de Rabanales, Universidad de Córdoba, 14071-Córdoba, Spain

Correspondence
Conrado Moreno-Vivián
bb1movic{at}uco.es

	ABSTRACT

TOP ABSTRACT Introduction Assimilatory and respiratory... Denitrification in archaea N2 fixation in archaea Transport and assimilation of... Regulation of nitrogen... Concluding remarks REFERENCES

 
The nitrogen cycle (N-cycle) in the biosphere, mainly driven by prokaryotes, involves different reductive or oxidative reactions used either for assimilatory purposes or in respiratory processes for energy conservation. As the N-cycle has important agricultural and environmental implications, bacterial nitrogen metabolism has become a major research topic in recent years. Archaea are able to perform different reductive pathways of the N-cycle, including both assimilatory processes, such as nitrate assimilation and N₂ fixation, and dissimilatory reactions, such as nitrate respiration and denitrification. However, nitrogen metabolism is much less known in archaea than in bacteria. The availability of the complete genome sequences of several members of the eury- and crenarchaeota has enabled new approaches to the understanding of archaeal physiology and biochemistry, including metabolic reactions involving nitrogen compounds. Comparative studies reveal that significant differences exist in the structure and regulation of some enzymes involved in nitrogen metabolism in archaea, giving rise to important conclusions and new perspectives regarding the evolution, function and physiological relevance of the different N-cycle processes. This review discusses the advances that have been made in understanding nitrate reduction and other aspects of the inorganic nitrogen metabolism in archaea.

	Introduction

TOP ABSTRACT Introduction Assimilatory and respiratory... Denitrification in archaea N2 fixation in archaea Transport and assimilation of... Regulation of nitrogen... Concluding remarks REFERENCES

 
As nitrogen is an essential element for living organisms, the availability of a suitable nitrogen source often limits primary productivity in both natural environments and agriculture. Nitrogen can be found in several oxidation states, from +5 in the most oxidized compound (nitrate) to –3 in the most reduced form (ammonium), but in biological compounds it is almost exclusively present in the fully reduced state. The global biogeochemical nitrogen cycle (N-cycle) allows the interconversions of nitrogen compounds to maintain a relatively small amount of fixed or combined nitrogen in a continuous exchange with atmospheric dinitrogen (N₂), a huge nitrogen reservoir. The N-cycle includes both reductive and oxidative processes, in which prokaryotes play a predominant role (Fig. 1). Assimilatory pathways, such as N₂ fixation and nitrate assimilation, generate ammonium that is further incorporated into cell material, mainly by the glutamine synthetase–glutamate synthase (GS–GOGAT) route. However, whereas N₂ fixation is only driven by free-living or symbiotic diazotrophic prokaryotes, assimilatory nitrate reduction is carried out by bacteria, fungi, algae and higher plants, and almost all living organisms are able to incorporate ammonium into carbon skeletons. On the other hand, some prokaryotes can obtain metabolic energy by redox processes involving nitrogen compounds, such as nitrification and denitrification. Autotrophic nitrification is a two-step process consisting of the oxidative conversion of ammonia to nitrite via hydroxylamine, carried out by ammonia-oxidizing bacteria, and the further oxidation of nitrite to nitrate, performed by nitrite-oxidizing chemolithoautotrophic bacteria. Denitrification is a respiratory process whereby nitrate is successively reduced to nitrite, NO, N₂O and N₂, which takes place predominantly under anaerobic conditions in many facultative bacteria (Berks et al., 1995; Zumft, 1997; Moreno-Vivián & Ferguson, 1998; Moreno-Vivián et al., 1999; Richardson & Watmough, 1999). Some heterotrophic bacteria are able to combine nitrification and aerobic denitrification by oxidizing ammonia to nitrite, which is later reduced to N₂ (de Boer & Kowalchuk, 2001). Ammonification is the dissimilatory reduction of nitrate to ammonia that does not serve the purpose of nitrogen autotrophy, although this term normally describes the release of ammonia from organic nitrogen compounds (Zumft, 1997). Finally, anaerobic ammonium oxidation (anammox) is a reaction that produces N₂ by reducing nitrite and oxidizing ammonium. This process, performed by bacterial communities of planctomycetes, seems to be of ecological importance in marine environments (Dalsgaard et al., 2003; Kuypers et al., 2003).

View larger version (21K): [in this window] [in a new window]   Fig. 1. The biological N-cycle. The different nitrogen compounds are arranged according to their oxidation states and the main reductive or oxidative pathways are indicated by arrows. Conversion of hydroxylamine to nitrous oxide by heterotrophic nitrification, and production of N2 from ammonium and nitrite by anaerobic ammonium oxidation (anammox), are not indicated. The enzymes involved in the different pathways are: 1–2, glutamine synthetase–glutamate synthase (GS–GOGAT); 3, glutamate dehydrogenase (GDH); 4, nitrogenase; 5, assimilatory nitrate reductase (Nas); 6, assimilatory nitrite reductase (sirohaem-Nir); 7, dissimilatory and respiratory nitrate reductases (Nap and Nar); 8, respiratory nitrite reductases (Cu-Nir and cd1-Nir); 9, nitric oxide reductase (Nor); 10, nitrous oxide reductase (Nos); 11, ammonia monooxygenase; 12, hydroxylamine oxidase; 13, nitrite oxidase.

	Assimilatory and respiratory nitrate reduction in archaea

TOP ABSTRACT Introduction Assimilatory and respiratory... Denitrification in archaea N2 fixation in archaea Transport and assimilation of... Regulation of nitrogen... Concluding remarks REFERENCES

 
Many archaea are able to reduce nitrate by assimilatory or respiratory pathways, dissimilatory nitrate reduction being much more frequent than nitrate assimilation (Martínez-Espinosa et al., 2001b). In addition, genes encoding putative nitrate transporters, nitrate reductases and nitrite reductases have been found in the sequenced genomes of members of both crenarchaeota and euryarchaeota (Table 1). However, in most cases, either functional or sequence data, but not both, are available. Therefore, future research should be directed to characterizing at the physiological and biochemical levels the products of the genes identified by these genome sequencing projects and their regulation. This will be of special evolutionary interest because significant differences between the bacterial and archaeal nitrate-reducing systems are emerging.

Preliminary sequence comparisons of Nar, Nas and Nap indicated that these enzymes represent different phylogenetic groups (Reyes et al., 1996). A recent study of the sequences of eukaryotic and prokaryotic nitrate reductases revealed the existence of three different clades, corresponding to the eukaryotic nitrate reductase, the prokaryotic respiratory Nar, and the prokaryotic Nas/Nap enzymes, with the Nap from Desulfovibrio desulfuricans providing the evolutionary link between Nas and Nap subclades (Stolz & Basu, 2002). It is worth noting from this study that the deepest branch of the Nar clade is the euryarchaeote Haloarcula marismortui, and that the Nar enzymes of the thermophilic bacterium Thermus thermophilus and the crenarchaeote Aeropyrum pernix are closely related. Phylogenetic trees based on NarH sequences also correlate well with the 16S rDNA tree, making the archaea A. pernix and Pyrobaculum aerophylum the outgroup and closely joining these archaeal sequences to the T. thermophilus bacterial subtree (Petri & Imhoff, 2000; Philippot, 2002). However, the NarG-based phylogeny is not entirely consistent with the 16S rDNA-based taxonomy (Chèneby et al., 2003; Gregory et al., 2003). As the nar genes were present before the phylogenetic divergence of bacteria and archaea, it can be assumed that respiratory nitrate reductase played a key role in energy metabolism during pre-oxic times (Petri & Imhoff, 2000). In addition, there is evidence that Nap and Nar have probably been acquired in some bacteria by horizontal gene transfer (Stolz & Basu, 2002). This seems to be the case for the nar gene cluster of T. thermophilus (Ramírez-Arcos et al., 1998b) and for the plasmid-encoded nap genes of Rhodobacter sphaeroides (Castillo et al., 1996). Results of NarG-based trees are also consistent with the possibility of lateral gene transfer (Gregory et al., 2003). Evidence of recent lateral gene transfer among hyperthermophilic archaea has been published (DiRuggiero et al., 2000), but the relative importance of this mechanism of genetic exchange among archaea, and also from and to bacteria, should be determined in future studies.

Nitrate reductase cofactors
Genome sequences of several archaea contain MGD cofactor biosynthesis genes, and archaeal nitrate reductases seem to be molybdoenzymes. However, tungsten and molybdenum have similar chemistry and it is known that hyperthermophilic archaea possess W-containing enzymes, such as aldehyde ferredoxin oxidoreductase, with the basic structure of the pterin cofactor shared in the Mo- and W-enzymes (Hille, 2002). Recently, Mo-free nitrate reductases with vanadium and haem c as cofactors have been isolated from vanadate-reducing, iron-reducing and sulfur-oxidizing bacteria (Antipov et al., 1998, 2003; Murillo et al., 1999). In addition, active W-substituted molybdoenzymes have been described in mesophilic bacteria (Buc et al., 1999; Gates et al., 2003). Thus, tungstate can function at the active site of the Paracoccus pantotrophus NapA, although the enzyme shows lower affinity for nitrate (Gates et al., 2003). Although nitrate reductases and other Mo-enzymes are usually inactivated by tungstate, this inhibition is not relevant in vivo since bacteria exhibit high-affinity Mo-uptake systems for extracting traces (micromolar or nanomolar) of Mo from the media, and because Mo is usually present in natural environments at higher concentrations than W. However, Mo(IV) forms a highly insoluble sulfide (MoS₂), while V (as VO²⁺ or V³⁺) and W (as or ) form relatively soluble salts, and therefore there are chemical reasons to believe that both W and V were more available than Mo in the primitive strongly reducing sulfide-rich environments (Williams & Fraústo da Silva, 2002). In addition, some extreme environments such as hot springs and deep-sea hydrothermal vents usually contain high levels of tungsten, and in fact an obligate W requirement for the anaerobic growth of the hyperthermophile Pyrobaculum aerophilum has been reported (Afshar et al., 1998). The respiratory nitrate reductase of this archaeon is a Mo-enzyme, but its activity is not abolished at high W concentrations, in contrast to bacterial Nar systems. Moreover, the presence of tungstate in the culture medium stimulates the anaerobic growth with nitrate (Völkl et al., 1993). The ability to respire nitrate in these conditions has been interpreted as an adaptation of the archaeon to the W-rich hyperthermophilic environment (Afshar et al., 1998, 2001). Clearly, further biochemical studies will be necessary to determine the possible involvement of alternative metal cofactors in archaeal nitrate reduction, but the possibility that V or W, rather than Mo, may be included in some archaeal nitrate reductase cofactors cannot be excluded.

Nitrate transport systems
Two types of nitrate transporters are involved in bacterial assimilatory nitrate reduction: the ATP-dependent ABC transporters composed of an integral membrane subunit, a cytoplasmic ATP-binding component and a periplasmic substrate-binding protein; and the monomeric NarK-type transporters belonging to the major facilitator superfamily, which depend on proton-motive force. However, only transport systems of the NarK type are involved in bacterial respiratory nitrate reduction (Moir & Wood, 2001). Within the bacterial NarK-like transporters there are two distinct subgroups. Recent studies in the bacterium Paracoccus pantotrophus suggest that NarK1 is a proton : nitrate symporter that allows initiation of nitrate respiration, whereas NarK2 is a nitrate : nitrite antiporter required for maintenance of a steady-state rate (Wood et al., 2002). Genome sequencing projects reveal that ABC-type transporters are widespread in archaea (Table 1). In particular, putative ABC-type nitrate transporters are present in the crenarchaeote Pyrobaculum aerophilum (Fitz-Gibbon et al., 2002) and the euryarchaeota Archaeoglobus fulgidus (Klenk et al., 1997), Methanocaldococcus jannaschii (Bult et al., 1996), Methanopyrus kandleri (Slesarev et al., 2002), Thermoplasma acidophilum (Ruepp et al., 2000) and Thermoplasma volcanium (Kawashima et al., 2000). However, substrate specificity for these putative transporters has been assigned solely on the basis of sequence similarities. A functional characterization of archaeal nitrate transporter genes has been only performed in Haloferax volcanii, where three ABC-type transporters are essential for nitrate respiration (Wanner & Soppa, 1999). The type of transport system used for nitrate uptake has a serious impact on the bioenergetics of the process because an ATP-driven nitrate uptake system is equivalent to consuming about three protons and this energetic cost makes its involvement in dissimilatory processes unlikely (Moir & Wood, 2001). For this reason, as mentioned above, bacterial nitrate assimilation usually requires ATP-dependent ABC nitrate transporters whereas nitrate respiration is associated with proton-motive-force-driven NarK transporters (Moreno-Vivián et al., 1999; Moir & Wood, 2001), even in the extreme thermophilic bacterium Thermus thermophilus (Ramírez et al., 2000). It is worth noting that most archaea with putative ABC nitrate transporters seem to contain respiratory nitrate reductases rather than assimilatory Nas enzymes. However, as discussed below, nitrate uptake may not be required for respiratory nitrate reduction in archaea, in contrast to bacteria, and the putative ATP-dependent nitrate transporters may indeed be involved in nitrate assimilation.

Assimilatory nitrate reduction in archaea
Most nitrate-reducing archaea use nitrate as alternative electron acceptor in anaerobic respiration, but recently, assimilatory nitrate and nitrite reductases have been purified from Haloferax mediterranei (Martínez-Espinosa et al., 2001a, b). This extreme halophilic archaeon grows aerobically with nitrate as the sole nitrogen source, but it can also denitrify because it contains both Nas and Nar enzymes. Assimilatory nitrate reductase is a dimer of 105 and 50 kDa subunits and uses ferredoxin as reductant, but not NAD(P)H (Martínez-Espinosa et al., 2001b). This could indicate a different structure for archaeal Nas because bacterial ferredoxin-Nas enzymes are usually monomeric, whereas NADH-dependent Nas enzymes are heterodimers (Moreno-Vivián et al., 1999). The assimilatory enzyme of H. mediterranei has a K_m for nitrate of 0·95 mM and an optimum temperature of 80 °C at 3·1 M NaCl, but 60 °C at 1·3 M NaCl. Activity is induced by nitrate and repressed by ammonium, as described for the bacterial Nas (Martínez-Espinosa et al., 2001b). The assimilatory nitrite reductase of H. mediterranei has also been purified. It is a 66 kDa monomer which uses ferredoxin as electron donor, has a K_m for nitrite of 8·6 mM, and shows maximal activity at 60 °C with 3·3 M NaCl. Like the bacterial assimilatory nitrite reductases, the H. mediterranei enzyme contains sirohaem and Fe–S centres (Martínez-Espinosa et al., 2001a). The 21 kDa ferredoxin from this halophilic archaeon has been purified and characterized, and its physiological role as the electron donor for both assimilatory nitrate and nitrite reductases has been confirmed (Martínez-Espinosa et al., 2003).

Analysis of the published archaeal genome sequences suggests the presence of putative assimilatory nitrate/nitrite reductases (Table 1). Thus Methanothermobacter thermautotrophicus has a putative nitrate reductase (MTH1567 gene product) homologous to bacterial Nas (Smith et al., 1997). Thermoplasma strains contain nitrate reductase related proteins (T. acidophilum Ta0136 and T. volcanium VN0215 gene products), which are similar to each other and to the eukaryotic assimilatory NADPH-nitrate reductases (Kawashima et al., 2000; Ruepp et al., 2000), although these putative proteins are too small (about 200 residues) compared to the nia-encoded eukaryotic nitrate reductases. The eukaryotic nitrate reductases are quite distinct from any of the bacterial Nas, Nar or Nap enzymes. They share no sequence similarity and do not bind the Mo-bisMGD cofactor, but instead have the so-called molybdopterin-cofactor (Mo-co), in addition to FAD and cytochrome b₅₅₇ (Moreno-Vivián et al., 1999; Campbell, 2001). Therefore, the assignation of a role in nitrate reduction to the Thermoplasma small nitrate-reductase-related proteins will be of special evolutionary interest and should be explored in future studies.

The Pyrobaculum aerophilum genome contains a gene (nirA) encoding a putative assimilatory ferredoxin-dependent sirohaem nitrite reductase (PAE2577), in addition to putative genes encoding a respiratory cytochrome cd₁-nitrite reductase (Fitz-Gibbon et al., 2002). Putative nirA homologous genes are also present in the genomes of Pyrococcus furiosus, Pyrococcus horikoshii (PF1197 and PH0827, respectively; Maeder et al., 1999), Aeropyrum pernix (APE1326; Kawarabayasi et al., 1999), Archaeoglobus fulgidus (AF0164; Klenk et al., 1997), Methanocaldococcus jannaschii (MJ0551; Bult et al., 1996), and Methanopyrus kandleri (MK0801; Slesarev et al., 2002). However, biochemical and functional studies are necessary to confirm the assimilatory nature of these putative nitrate and nitrite reductases and to establish if these archaeal enzymes are structurally different from the bacterial assimilatory enzymes.

Respiratory nitrate reduction in archaea
The ability to use nitrate as a terminal electron acceptor in energy metabolism is found in a variety of halophilic and hyperthermophilic archaea; many of them can perform the entire denitrification process (see below). Respiratory Nar enzymes have been purified from several denitrifying halophilic euryarchaeota, including three Haloferax species and Haloarcula marismortui (Table 1). A salt-requiring Nar has been purified from Haloferax mediterranei, with an activity maximum at 89 °C in 3·2 M NaCl and K_m for nitrate values in the range from 2·5 to 6·7 mM depending on salt concentration (Álvarez-Ossorio et al., 1992). The membrane-bound Nar of Haloferax denitrificans is composed of two subunits of 116 and 60 kDa, with a K_m for nitrate of 0·2 mM. Curiously, the enzyme is stable in the absence of salt, and activity decreases with increasing NaCl concentrations (Hochstein & Lang, 1991). Haloferax volcanii contains a membrane-bound trimeric Nar, with 100, 61 and 31 kDa subunits. Activity is salt-independent, with a temperature optimum of 80 °C and a K_m for nitrate of 0·36 mM (Bickel-Sandkötter & Ufer, 1995). The Haloarcula marismortui Nar has a K_m for nitrate of 80 µM with 2·0 M NaCl, and activity is enhanced by salt. The enzyme was first described as a homotetramer of a 63 kDa polypeptide (Yoshimatsu et al., 2000), but further characterization revealed that it is a heterodimer with 117 and 47 kDa subunits, the 63 kDa band being an incomplete denaturation state of the enzyme (Yoshimatsu et al., 2002). It has been proposed that the H. marismortui Nar is a new archaeal type of membrane-bound nitrate reductase (Yoshimatsu et al., 2002). First, the enzyme is composed of two subunits, which are homologous to the NarGH bacterial subunits, but lacks the NarI membrane-associated protein. Second, the H. marismortui NarGH complex has significant sequence and structure similarity to the dissimilatory selenate reductase from Thauera selenatis, although it does not reduce selenate. Recently, the similarity of Nar to selenate reductases has also been indicated on the basis that both enzymes have Asp ligands to the Mo atom (Jormakka et al., 2004). In addition, the N-terminal region of NarG includes a typical twin-arginine signal peptide for protein translocation across the membrane by the Tat export pathway (Berks et al., 2000) and the enzyme has activity in situ with both membrane-permeable benzyl viologen and membrane-impermeable methyl viologen, suggesting that the catalytic site is located in the outside of the membrane, as described for the T. selenatis selenate reductase (Yoshimatsu et al., 2002). In contrast, bacterial NarG faces the cytoplasmic side of the membrane and nitrate reduction takes places in the cytoplasm (Berks et al., 1995; Moreno-Vivián et al., 1999). Finally, although the H. marismortui NarGH complex does not contain any quinol-oxidizing cytochrome b subunit, a gene encoding a homologue to SerC, the cytochrome b subunit of selenate reductase, is found downstream of the narGH genes and together with a narJ-homologous gene, which probably encodes a nitrate-reductase-specific chaperone for MGD cofactor assembly.

A gene cluster encoding a putative respiratory Nar has also been identified in Archaeoglobus fulgidus (Klenk et al., 1997; Table 1). The AF0176 gene encodes a protein with 42 % similarity to Escherichia coli NarG, although this A. fulgidus NarG protein is smaller (80 kDa) and contains a twin-arginine signal peptide, suggesting that it is translocated across the membrane like the H. marismortui and probably other archaeal NarG proteins. The AF0175 product is predicted to bind four Fe–S clusters, whereas the AF0174 gene encodes an integral membrane protein, although no haem groups are bound to this protein. Finally, a NarJ homologue is also found in the gene cluster (Klenk et al., 1997; Richardson et al., 2001). In addition, genes encoding an ABC-type nitrate transporter (AF0638–AF0640), a putative nirA gene encoding the cytoplasmic sirohaem nitrite reductase (AF0164) and other putative iron–sulfur proteins clustered to narI-like genes (AF0499–AF0501 and AF0543–AF0547) are also present in the A. fulgidus genome. Some hyperthermophilic archaea are also able to respire nitrate. Thus, Aeropyrum pernix has a putative nar cluster (Kawarabayasi et al., 1999) and Pyrobaculum aerophilum uses nitrate as respiratory electron acceptor (Völkl et al., 1993). The purified Nar of this archaeon shows very high specific activity, with a K_m for nitrate of 58 µM and an optimal temperature of 95 °C, and is active in the presence of high tungstate concentrations. Interestingly, the P. aerophilum Nar is a heterotrimer, with subunits of 130, 52 and 32 kDa, and contains a b-type cytochrome similar to the Thauera selenatis SerC, in addition to molybdenum cofactor and iron–sulfur cluster (Afshar et al., 2001).

Bacterial and archaeal Nar systems also differ in the organization of the nar genes (Philippot, 2002). The main narGHJI operon organization is conserved in all bacteria, including Thermus thermophilus (Ramírez-Arcos et al., 1998a), although this organism has an extra narC gene in the operon, and Paracoccus pantotrophus includes the narK gene in the narKGHJI cluster (Wood et al., 2001). However, archaeal nar genes do not conserve this organization. Thus, in Pyrobaculum aerophilum a putative narJ gene is located upstream from the narGH genes, in the opposite direction. A gene encoding a cytochrome b is present downstream from narH, although its role as Nar subunit is conjectural, and another putative narI homologue is located elsewhere in the genome (Fitz-Gibbon et al., 2002). In Haloarcula marismortui, two small ORFs and a narJ homologue are located downstream from the narGH genes, and putative genes encoding an iron–sulfur protein and a cytochrome b are located upstream from narGH (Yoshimatsu et al., 2002). In Aeropyrum pernix, no narI-like gene has been found and narGH genes are followed by a putative cytochrome b gene and narJ, which is located about 2 kb downstream from narH (Kawarabayasi et al., 1999). Therefore, based on subunit composition, subcellular location of the active site and nar gene organization, it can be concluded that archaeal Nars are a new type of enzyme with the active site facing the outside and anchored to the membrane by a SerC-like cytochrome b, or stabilized by the lipid environment in the membrane as described for the P. aerophilum Nar (Afshar et al., 2001). This system could be an ancient respiratory nitrate reductase, although the nitrite formed could also be assimilated (Richardson et al., 2001). Figs 2 and 3 schematize the different organization of the bacterial and archaeal respiratory nitrate-reducing systems, and the other enzymes involved in the whole denitrification pathway (see below).

View larger version (34K): [in this window] [in a new window]   Fig. 2. Schematic representation of the denitrification pathway in bacteria. The periplasmic nitrate reductase (NAP), the membrane-bound nitrate reductase (NAR), the periplasmic nitrite reductases (Cu-NIR or cd1-NIR), the membrane-bound nitric oxide reductase (NOR) and the periplasmic nitrous oxide reductase (NOS) are represented without indicating their subunit composition and cofactors. Note that not all represented components are present in a single organism (i.e. NAP and NAR, Cu-NIR and cd1-NIR). cyt bc1, proton-pumping cytochrome bc1 complex; cyt c, NapC membrane-bound tetrahaem cytochrome c; cyt c550, cytochrome c550; MQH2/MQ, menaquinol/menaquinone pool.

View larger version (23K): [in this window] [in a new window]   Fig. 3. Schematic representation of the denitrification pathway in archaea. Membrane location and dependence on menaquinol as the electron donor are assumed for the four reductases (nitrate reductase, NAR; nitrite reductase, NIR; nitric oxide reductase, NOR; and nitrous oxide reductase, NOS), as suggested for Pyrobaculum aerophilum. These reductases are represented without indicating their subunit composition and cofactors. An outside location for the nitrate reduction site in the NAR complex is also assumed (see text for details). The identity of the proton-translocating dehydrogenase (DH) that regenerates menaquinol from menaquinone is presently unknown. The putative ABC-type nitrate/nitrite transporter present in some archaea is assumed to serve for assimilatory purposes rather than for the respiratory nitrate reduction, since the active site of NAR is assumed to be located on the outside of the membrane. MQH2/MQ, menaquinol/menaquinone pool.

	Denitrification in archaea

TOP ABSTRACT Introduction Assimilatory and respiratory... Denitrification in archaea N2 fixation in archaea Transport and assimilation of... Regulation of nitrogen... Concluding remarks REFERENCES

 
Denitrification can be considered as the modular assemblage of four partly independent respiratory processes: nitrate, nitrite, nitric oxide and nitrous oxide reduction (Zumft, 1997; Figs 2 and 3). All denitrifiers isolated so far are also able to respire oxygen, indicating a close evolutionary relationship between the two processes, and denitrification has been considered the ancestor of aerobic respiration (Saraste & Castresana, 1994). Therefore, knowledge of denitrification in archaea is of special interest from an evolutionary point of view. Denitrification has been described for several halophilic archaea, such as Haloferax and Haloarcula strains (Tomlinson et al., 1986; Inatomi & Hochstein, 1996; Ichiki et al., 2001), and extreme thermophilic archaea, such as Ferroglobus placidus (Vorholt et al., 1997) and Pyrobaculum aerophilum (Völkl et al., 1993; de Vries & Schröder, 2002). As the archaeal respiratory nitrate reductases that catalyse the first step of denitrification have been described above, we will focus now on the nitrite and N-oxide respiration processes in archaea.

Nitrite respiration
In bacteria, there are two different and unrelated NO-producing respiratory nitrite reductases: the homotrimeric copper-containing enzyme encoded by nirK, and the homodimeric cytochrome cd₁-nitrite reductase encoded by nirS. The crystal structures of both enzymes, which are located in the periplasm, are known and have been reviewed in detail elsewhere (Zumft, 1997; Ferguson, 1998; Richardson & Watmough, 1999). Azurin and pseudoazurin are the electron donors to Cu-Nir, whereas cytochrome c₅₅₁ is the reductant for cd₁-Nir (Zumft, 1997).

Denitrifying archaea also contain either Cu-Nir or cd₁-Nir. Haloarcula marismortui has a nirK gene encoding a Cu-Nir. EPR analysis reveals the presence of type 1 and type 2 Cu centres in the purified enzyme, which has optimal activity at 2 M NaCl. A 46 kDa precursor of this protein with an N-terminal signal peptide for translocation across the plasma membrane has been detected by SDS-PAGE (Ichiki et al., 2001). Although the physiological electron donor is unknown, the archaeal halocyanin, a blue Cu-protein present in some archaea (Scharf & Engelhard, 1993), could be the putative reductant. Protein sequence analysis of the H. marismortui nirK gene product reveals a close relationship with the Neisseria gonorrhoeae Cu-Nir, thus suggesting lateral gene transfer of the nirK gene (Ichiki et al., 2001; Philippot, 2002). Interestingly, the Haloferax denitrificans nirK gene sequence is also available (CAD89521, Table 1), and is highly similar to both H. marismortui and N. gonorrhoeae Cu-Nir. However, antiserum against bacterial Cu-Nir does not recognize the H. denitrificans enzyme, suggesting possible structural differences (Inatomi & Hochstein, 1996).

The existence of a nirS homologous gene in the genome of the hyperthermophile Pyrobaculum aerophilum suggests the presence of a cd₁-nitrite reductase (Table 1). Recently, it has been reported that this archaeon has a membrane-bound respiratory nitrite reductase which uses menaquinol as the physiological electron donor, thus having different subcellular location and reductant than bacterial cd₁-Nir (Figs 2 and 3). It is worth noting that, in contrast to denitrifying bacteria, all four reductases involved in denitrification in P. aerophilum are membrane-bound enzymes which use menaquinol as electron donor (de Vries & Schröder, 2002). The dependence of the menaquinol pool and the membrane confinement of all these enzymes in archaeal denitrification could be related to the relatively small volume between the S-layer and the cytoplasmic membrane, in comparison with the larger periplasmic space in bacteria, or could reflect the ancient origin of this pathway. The membrane location of all denitrification enzymes in P. aerophilum could also explain the absence in this archaeon of soluble components of the denitrification pathway, such as c-type cytochromes or cupredoxins (Lubben & Morand, 1994; de Vries & Schröder, 2002; see Figs 2 and 3).

N-oxide respiration
Nitric oxide, the product of the respiratory nitrite reductases, is a toxic compound that is reduced to N₂O by nitric oxide reductases (Nor) immediately after it has been generated. Several enzymes with Nor activity have been described. In fungal denitrification, NO reduction to N₂O is catalysed by a soluble monomeric enzyme of the cytochrome P-450 family (Shoun & Tanimoto, 1991; Nakahara et al., 1993). In most denitrifying bacteria, Nor is a membrane complex of a 17 kDa cytochrome c (encoded by norC) and a 38 kDa cytochrome b with 12 transmembrane regions (encoded by norB). This enzyme, known as cNor, receives electrons from cytochrome c. The denitrifying bacterium Ralstonia eutropha and several non-denitrifying pathogenic bacteria such as Neisseria species possess a monomeric Nor with 14 transmembrane regions, which is called qNor because it is a quinol-oxidizing enzyme. This protein, encoded by norZ, is similar to the norB gene product but contains an N-terminal extension absent in NorB, with a quinone-binding site. In Gram-positive bacteria, a menaquinol Nor with an additional small Cu_A-containing subunit has been described as qCu_ANor (Hendriks et al., 2000; de Vries & Schröder, 2002; Philippot, 2002).

Genetic and biochemical evidence indicates that the archaeon Pyrobaculum aerophilum has a qNor-type NO reductase (Table 1). Putative nir (nitrite reductase) and nor (nitric oxide reductase) genes are located 3·3 kb upstream from the nar (nitrate reductase) genes (Völkl et al., 1993; Fitz-Gibbon et al., 2002). The purified Nor is a single membrane subunit with two modified haem groups and one non-haem iron, and uses menaquinol as electron donor (de Vries & Schröder, 2002). The genome of Methanosarcina mazei also contains a NO reductase gene (MM2988; Deppenmeier et al., 2002) and Sulfolobus solfataricus includes a putative gene encoding a haem NO reductase, but it is truncated by a transposase separating the N-terminal (SSO1573) and C-terminal (SSO1571) domains (She et al., 2001). Sequenced genomes of other archaea do not include nor homologous genes (Table 1), although the presence of NO reductase has been reported in some denitrifying halophilic archaea that have not yet been sequenced. The presence of qNor in some archaea is of evolutionary interest because qNor could be the ancestor not only of the cNor, by acquiring haem c-binding residues in the N-terminal region and separating this domain as a small cytochrome c subunit, but also of the whole superfamily of haem-copper oxidases and aerobic respiratory enzymes. The capacity for N-oxide respiration appeared early in evolution and qNor may have arisen to detoxify NO; this could still be the role of the enzyme in non-denitrifying bacteria. After the appearance in the atmosphere of oxygen generated photosynthetically, these enzymes could have replaced non-haem iron by copper and evolved to generate the proton-pumping cytochrome oxidases (Saraste, 1994; Saraste & Castresana, 1994; Richardson, 2000; de Vries & Schröder, 2002).

On the other hand, the E. coli flavorubredoxin has been recently described as a novel type of prokaryotic NO reductase. The purified protein has a turnover number similar to that of canonical bacterial Nor and also has oxygen reductase activity, although it shows lower affinity for O₂ than for NO (Gomes et al., 2002). Flavorubredoxins are A-type flavoproteins which are present in bacteria and some archaea, such as Pyrococcus furiosus, Archaeoglobus fulgidus, Methanocaldococcus jannaschii and Methanothermobacter thermautotrophicus. Therefore, it has been proposed that this protein is a widespread class of NO reductase, which could detoxify NO or oxygen, allowing versatility in response to environmental conditions (Gomes et al., 2002). However, the physiological relevance of this protein should be determined in future studies.

The last step of denitrification, the reduction of N₂O to N₂, is of great environmental importance because it closes the N-cycle. Although N₂O is less toxic than NO or nitrite, and bacteria could manage without converting N₂O into N₂, most denitrifying bacteria contain nitrous oxide reductases (Nos) encoded by the nosZ gene. These enzymes are periplasmic multicopper homodimers that receive electrons from cytochrome c or pseudoazurin (Zumft, 1997). In each 65 kDa monomer, the copper ions are organized in two centres, a di-copper cluster CuA resembling that of cytochrome oxidase, and a CuZ cluster in which four Cu atoms are ligated by seven His residues. A bridging inorganic sulfur in the CuZ centre has also been identified (Brown et al., 2000; Rasmussen et al., 2000). In some archaea, such as Haloarcula marismortui and Ferroglobus placidus, the predominant gas species produced by denitrification is N₂O (Werber & Mevarech, 1978; Vorholt et al., 1997), although other archaea such as Haloferax denitrificans and Pyrobaculum aerophilum are able to reduce N₂O to N₂, having a complete denitrification pathway (Tomlinson et al., 1986; Völkl et al., 1993; de Vries & Schröder, 2002). However, in P. aerophilum, N₂O formed by qNor accumulates because its reduction to N₂ by the nitrous oxide reductase is very slow. In contrast to the periplasmic bacterial Nos, the P. aerophilum enzyme is located in the membrane and uses menaquinol as electron donor, like the other three reductases of the denitrification pathway in this hyperthermophile (de Vries & Schröder, 2002; Fig. 3). N₂O reduction has not been investigated in other species of archaea and putative nos genes are not present in the sequenced archaeal genomes (Table 1), although two long hypothetical quinol oxidases (ST2395 and ST0105) with the motif of the Nos binuclear copper centre are found in Sulfolobus tokodaii (Kawarabayasi et al., 2001).

	N2 fixation in archaea

TOP ABSTRACT Introduction Assimilatory and respiratory... Denitrification in archaea N2 fixation in archaea Transport and assimilation of... Regulation of nitrogen... Concluding remarks REFERENCES

 
N₂ is the most abundant gas in Earth's atmosphere, but it is highly stable and unreactive and is only used as nitrogen source by some prokaryotic organisms. Biological N₂ fixation is catalysed by nitrogenase, an O₂-sensitive enzyme complex which generates two molecules of NH₃ by reduction of N₂. Nitrogenase can also reduce protons to H₂ and several substrates with double or triple bonds, such as acetylene and cyanide (Halbleib & Ludden, 2000). Nitrogenase consists of a molybdenum–iron protein (dinitrogenase), which receives electrons from an iron protein known as dinitrogenase reductase. MoFe protein is an ₂₂ tetramer encoded by the nifD and nifK genes, and Fe protein is a homodimer of the nifH gene product (Howard & Rees, 1996). Several nif genes are also involved in cofactor biosynthesis, electron transfer and regulation. Alternative nitrogenases have been found in several bacteria, with V or Fe substituting for Mo. These enzymes have lower catalytic efficiency than Mo-nitrogenases and are encoded by nif homologous genes, termed vnf (V-nitrogenase) or anf (Fe-nitrogenase). Alternative nitrogenases include an additional small subunit , encoded by vnfG or anfG (Eady, 1996; Rees, 2002). N₂ fixation is highly regulated since is an energy-expensive process requiring large amounts of reducing power and ATP. Ammonium causes a reversible short-term inhibition of enzyme activity, called switch-off/on (Zumft & Castillo, 1978), and a long-term effect at the transcriptional level. Ammonium switch-off occurs by ADP-ribosylation of NifH by DraT/DraG, an ADP-ribosyltransferase/glycohydrolase system. Expression of nif genes requires NifA, a transcriptional activator under the control of Ntr, a two-component regulatory system of bacterial nitrogen metabolism (Ludden, 1994; Merrick & Edwards, 1995; Halbleib & Ludden, 2000).

N₂ fixation in archaea, first described in Methanosarcina barkeri (Murray & Zinder, 1984) and Methanococcus thermolithotrophicus (Belay et al., 1984), is exclusive of, but widespread within, methanogenic euryarchaeota (Table 2). However, neither Methanococcus voltae nor Methanocaldococcus jannaschii grows diazotrophically, despite the presence of nifH-like genes in their genomes (Leigh, 2000). Mo-nitrogenase predominates in diazotrophic methanogens, which contain at least the nifHDKENX genes (Lobo & Zinder, 1990; Chien & Zinder, 1996; Kessler et al., 1998; Leigh, 2000). nifHDK are the structural genes for nitrogenase, whereas nifENX code for FeMo-cofactor biosynthesis proteins (Moreno-Vivián et al., 1989). The diazotrophic growth of Methanococcus maripaludis shows Mo dependence, indicating that N₂ fixation is catalysed by a Mo-nitrogenase (Kessler et al., 1997), although Methanosarcina species contain alternative nitrogenases (Table 2). M. barkeri has two clusters of nitrogenase genes, the nif2 gene cluster encoding the Mo-nitrogenase and the nif1 (vnf) cluster encoding an alternative V-nitrogenase, which also includes a vnfG gene (Chien et al., 2000). Like the nif regions of other methanogens, both gene clusters contain two glnB genes encoding nitrogen signal transduction proteins (Chien et al., 2000). The M. acetivorans genome includes nif, vnf and anf genes, being unique among archaea in having all three nitrogenase systems (Galagan et al., 2002).

Phylogenetic analysis suggests that N₂ fixation may have originated in a common ancestor of bacteria and archaea, before branching of these domains and prior to the appearance of oxygen in the atmosphere (Fani et al., 2000; Leigh 2000; Berman-Frank et al., 2003). Comparative analysis of the amino acid sequences of the NifDK and NifEN proteins suggests a two-step in-tandem duplication of an ancestor gene, giving rise first to a bicistronic operon, which later underwent a paralogous duplication leading to the nifDK and nifEN genes before the divergence of archaea and bacteria (Fani et al., 2000). The function of this ancient enzyme might have depended on the composition of the early atmosphere, being either a nitrogenase for N₂ fixation in a neutral anoxic atmosphere containing N₂ but not NH₃, or alternatively, a cyanide-detoxifying enzyme in a strongly reducing atmosphere, which could change its specificity to N₂, another triple-bond substrate, after progressive exhaustion of combined nitrogen, thus enabling survival of ancient micro-organisms in N-deficient environments before the appearance of photosynthetic O₂ (Fani et al., 2000; Berman-Frank et al., 2003). After the divergence of archaea and bacteria, many organisms lost nif genes due to the weakening of selection pressure and the high energetic cost of N₂ fixation, although lateral gene transfer of nif genes might have also occurred (Fani et al., 2000). However, there are no apparent reasons to explain the lost of diazotrophy in all archaea except methanogens. Earth's earliest organisms may have been similar to the thermophiles found in deep-sea hydrothermal vents, but currently the most thermophilic micro-organism known, Methanococcus thermolithotrophicus, is able to fix N₂ at 64 °C (Mehta et al., 2003). This makes especially interesting a recent paper describing the first evidence of potential nitrogen fixers in diffuse hydrothermal vent fluids and deep-sea water, and reporting that all of the nifH genes from these samples are most closely related to M. thermolithotrophicus (Mehta et al., 2003).

	Transport and assimilation of ammonium in archaea

TOP ABSTRACT Introduction Assimilatory and respiratory... Denitrification in archaea N2 fixation in archaea Transport and assimilation of... Regulation of nitrogen... Concluding remarks REFERENCES

 
Glutamine synthetase–glutamate synthase (GS–GOGAT) and glutamate dehydrogenase (GDH) are the major pathways for ammonium assimilation. The GS–GOGAT route requires ATP but has a high affinity for ammonium, whereas GDH does not consume ATP but is less effective in cells growing in N-limited conditions. As these enzymes are essential in ammonia assimilation and amino acid metabolism, they are present in all three domains of life. Movement of ammonium across biological membranes is also an important process, and ammonium transporters (Amt) are found in bacteria, archaea and eukarya.

Ammonium transport
N₂ fixation and nitrate assimilation generate ammonium for incorporation into carbon skeletons inside the cells. However, when ammonium is available in the environment, organisms usually take up this ion by a specific transport system, inhibiting and/or repressing the reductive assimilatory pathways of the N-cycle. Although NH₃ diffuses through biological membranes, and this could represent a significant process in organisms growing in alkaline environments, ammonium transporters are found in both prokaryotic and eukaryotic organisms. Amt proteins are a family of ammonium and methylammonium transporters, which are integral membrane proteins of 45–50 kDa with 10–12 transmembrane helices (Thomas et al., 2000). It is widely accepted that Amt proteins mediate an active uniport which depends on proton-motive force, although it has been also proposed that Amt proteins facilitate diffusion of NH₃ across the plasma membrane bidirectionally (Soupene et al., 2002). The E. coli AmtB protein is essential for transport of ammonium and is able to complement a Saccharomyces cerevisiae mutant (Blakey et al., 2002). Putative ammonium transporters (AmtB) are also present in most archaeal genomes (Table 2). In bacteria and archaea, the amtB gene is invariably linked to the regulatory glnK gene, which encodes a GlnB-like protein (P_II protein). Both GlnK and AmtB proteins interact for regulation of the transport activity in response to nitrogen availability (Thomas et al., 2000; Coutts et al., 2002). In E. coli, AmtB is inactivated by formation of a membrane-bound complex with GlnK, a process regulated by uridylylation/deuridylylation of GlnK in response to the nitrogen status of the cells, and the AmtB activity is also required for GlnK deuridylylation. Therefore, AmtB could be considered as an ammonium sensor, and transport is also an integral part of the signal transduction cascade (Javelle et al., 2004).

The glutamine synthetase–glutamate synthase cycle
Glutamine synthetase (GS) is a ubiquitous enzyme with a central role in nitrogen metabolism. Three types of GS catalyse the ATP-dependent synthesis of glutamine from glutamate and ammonia. GSI, found in most prokaryotes, is a dodecameric protein composed of identical 50–56 kDa subunits arranged in two hexagonal rings. GSII is mainly present in eukaryotes and has an octameric structure, with identical subunits of 45–50 kDa. GSIII is only found in a few bacterial species, such as Bacteroides fragilis, and has a hexameric structure with identical subunits of about 83 kDa (Pesole et al., 1995; DiRuggiero & Robb, 1996; Eisenberg et al., 2000).

Putative glnA genes encoding GSI are present in archaeal genomes, sometimes in several copies (Table 2). Sequence analysis of the glnA gene reveals two main subdivisions: GSI-, including euryarchaeota, Thermotoga and low-G+C Gram-positive bacteria; and GSI-, with all other bacteria. Biochemical properties also indicate that GSs from archaea and Gram-positive bacteria are functionally similar (Bhatnagar et al., 1986). Lateral gene transfer from archaea to Thermotoga and Gram-positive bacteria could explain the close evolutionary relation between these GSs (Tiboni et al., 1993; Pesole et al., 1995). However, the crenarchaeote Sulfolobus solfataricus GS is not placed in either the GSI- or the GSI- division (Brown et al., 1994), although the Sulfolobus acidocaldarius glnA is more related to GSI-, despite its enzyme properties being more similar to GSI- (Yin et al., 1998). Most GSI- enzymes are regulated by adenylylation/deadenylylation, but GSI- enzymes are not controlled by this reversible mechanism. Purified GSs from Pyrococcus sp., S. acidocaldarius and Methanobacterium ivanovi are dodecamers of identical subunits that are not regulated by adenylylation (Bhatnagar et al., 1986; Rahman et al., 1997; Yin et al., 1998). Therefore, branching of GSI- and GSI- could precede the separation of archaea and bacteria.

The relevance of GS in ammonium assimilation varies depending on the archaeal species. Most hyperthermophilic archaea show very high glutamate dehydrogenase activity, suggesting that this enzyme is the major pathway for ammonium assimilation. In fact, Thermococcus profundus has no detectable GS activity (Kabayashi et al., 1995) and the GS of Pyrococcus sp. has low biosynthetic activity, suggesting that the reaction is biased towards glutamate production. Therefore, hyperthermophilic archaea seem not to use GS predominantly for ammonium assimilation. In contrast, the inability to obtain a glnA null mutant in Methanococcus maripaludis suggests that this archaeon cannot transport glutamine and that GS is an essential enzyme, even in the presence of alternative nitrogen sources (Cohen-Kupiec et al., 1999).

Glutamate synthase (glutamine : 2-oxoglutarate amidotransferase, GOGAT) is an iron–sulfur flavoprotein that catalyses the transfer of the amide group of glutamine to 2-oxoglutarate, yielding two molecules of glutamate. In bacteria, GOGAT is usually an NADPH-dependent enzyme arranged in an ₄₄ structure, with large