Biochemical Society Focused Meeting

Oxygen control of nitrogen oxide respiration, focusing on α-proteobacteria

James P. Shapleigh


Denitrification is generally considered to occur under micro-oxic or anoxic conditions. With this in mind, the physiological function and regulation of several steps in the denitrification of model α-proteobacteria are compared in the present review. Expression of the periplasmic nitrate reductase is quite variable, with this enzyme being maximally expressed under oxic conditions in some bacteria, but under micro-oxic conditions in others. Expression of nitrite and NO reductases in most denitrifiers is more tightly controlled, with expression only occurring under micro-oxic conditions. A possible exception to this may be Roseobacter denitrificans, but the physiological role of these enzymes under oxic conditions is uncertain.

  • Agrobacterium tumefaciens
  • denitrification
  • nitrate reductase
  • nitric oxide reductase
  • α-proteobacterium
  • Rhodobacter sphaeroides


The processes that together define the nitrogen cycle are largely mediated by bacteria [1]. Oxygen strongly influences the growth and physiology of bacteria catalysing reactions in the nitrogen cycle. For example, nitrification occurs under oxic conditions, whereas anaerobic ammonia oxidation (annamox), as its name indicates, occurs under anoxic conditions. Other processes, such as nitrogen fixation, are negatively affected by oxygen, but some organisms have developed strategies allowing nitrogen fixation to occur under oxic conditions [2]. Processes linked with nitrate reduction, such as denitrification or ammonification, occur mainly under anoxic conditions. The net result of these varying sensitivities is that flux through the various processes changes dramatically in response to shifts in oxygen.

Denitrification, the dissimilatory reduction of nitrate to nitrogen gas, is regarded as an anoxic or micro-oxic process [3]. Complete denitrification requires four enzymes: nitrate reductase, nitrite reductase (Nir), nitric oxide reductase (Nor) and nitrous oxide reductase (Nos). Interestingly, no strictly anaerobic denitrifier has ever been isolated. Since denitrifiers are facultative aerobes, this means that they must choose between oxygen and nitrate if both are available. Oxygen is preferred by most isolates, which is generally ascribed to the fact that more energy is gained from oxygen respiration than nitrogen oxide respiration. However, with the exception of Nos, none of the terminal nitrogen oxide reductases are oxygen-sensitive, making it possible that the denitrification enzymes could be used under oxic conditions [4].

Oxygen preference is based on its advantages during respiration. However, it has become obvious that in some cases nitrogen oxide reductases are not being used for respiration. A good example of this is provided by examining the nitrogen oxide reductases present in the four Rhodobacter sphaeroides strains that have had their genomes sequenced. Analysis of the denitrification potential has showed that strain 2.4.3 has all four reductases, strain KD131 is missing nitrate reductase but has the remaining three, strain 2.4.1 has nitrate reductase and Nor, whereas strain 2.4.9 has only Nor. The latter two strains and possibly KD131 are unlikely to use nitrogen oxide reductases for respiration under micro-oxic conditions. This suggests an alternative physiological role for these enzymes and, by extension, it is possible the bioenergetic rationale for their repression under oxic conditions is no longer valid. With this in mind, I want to briefly summarize our current understanding of the physiological roles and expression of nitrate reductase, Nir and Nor in a few α-proteobacterial denitrifiers.

Nap regulation in select α-proteobacteria

There are several different types of nitrate reductase [5]. Two dissimilatory forms, Nar and Nap, are located in the membrane and periplasm respectively. The assimilatory form, Nas, is found in the cytoplasm. Denitrifiers will contain Nap and/or Nar. The α-proteobacterial denitrifiers my laboratory has worked with, R. sphaeroides and Agrobacterium tumefaciens, contain Nap. A. tumefaciens also has Nas, whereas R. sphaeroides lacks this enzyme.

In R. sphaeroides strain 2.4.1, the nap operon is located on a plasmid [6]. Recent work has shown that expression of the nap operon increases significantly when cells are shifted from anoxic photosynthetic conditions to oxic conditions [7]. The medium used for these experiments was not supplemented with nitrate, suggesting that the expression of nap does not require nitrate. Expression under low oxygen conditions was enhanced in a mutant lacking PrrA, the response regulator of the two-component PrrBA system [8]. PrrBA has been shown to regulate a large number of genes [8]. Available evidence suggests that the phosphorylated form of the response regulator PrrA is responsible for repressing nap expression under low oxygen. Since this enzyme is maximally expressed when oxygen is present, it seems unlikely that it is being used for respiration.

In another R. sphaeroides strain, DSM158, a similar pattern of expression was observed [9]. Highest expression was observed under oxic conditions, and nitrate did not seem to be required for expression. However, there was no nitrate reductase activity under oxic conditions. Under photosynthetic conditions, there was evidence of nitrate reductase activity. Growth on the reduced carbon sources butyrate and caproate under these conditions was severely impaired in a strain lacking the nap cluster [9]. This dependency on Nap indicates that nitrate reductase activity helps dispose of excess electrons generated by catabolism of these electron-rich compounds. Recent work has shown that non-sulfur purple photosynthetic bacteria place a high demand on redox homoeostatic pathways during growth [10]. The use of Nap in redox homoeostasis has been shown previously in Paracoccus pantotrophus and its relatives [11,12].

A slightly different nap expression pattern was observed in R. sphaeroides strain IL106, which like 2.4.3 is a denitrifying strain [13]. Insertional inactivation of the genes in the nap cluster eliminated nitrate reductase activity, indicating that this is the only cluster of genes encoding a dissimilatory nitrate reductase in this strain. This nap gene cluster showed the same level of expression under both oxic and denitrifying conditions. There was also evidence that nitrate is an effector for expression. Expression increased in the presence of the reduced carbon compound butyrate, indicating that this protein probably plays a role in redox homoeostasis.

R. sphaeroides strain 2.4.3 has a nap cluster orthologous to the nap clusters in DSM158, IL106 and 2.4.1. As in 2.4.1, this nap is present on a plasmid. Unlike the other R. sphaeroides strains, 2.4.3 has a second nap cluster. This cluster is located on the smaller of the two chromosomes which has been shown to be evolving at a rapid rate due to lateral gene transfer [14]. This cluster is not paralogous to the other nap cluster in this strain, since it contains two additional genes napGH. nap operons with napGH are not commonly found in α-proteobacteria nor is it common for bacteria to have more than one nap cluster, but some species of Shewanella have also been found with more than one nap cluster [15]. Expression of the nap clusters of 2.4.3 will be reported elsewhere (A. Hartsock and J.R. Shapleigh, unpublished work).

A. tumefaciens has a nap cluster orthologous to the plasmid-borne version found in R. sphaeroides 2.4.3 and 2.4.1. This nap showed yet another pattern of expression, since it was maximally expressed under denitrifying conditions (Figure 1). As with 2.4.1, its expression was not influenced by nitrate. The regulator(s) responsible for expression of nap in A. tumefaciens are uncertain, since there are no conserved sequence motifs in the promoter region that might indicate the involvement of a particular family of regulators.

Figure 1 Expression of napF of A. tumefaciens under a variety of conditions measured using a napF–lacZ fusion

Growth of cells and assay were as described previously [20]. Measurements were done in triplicate and the S.D. is <10%. Oxic early refers to growth under oxic conditions with cells at a D600 of 0.25. Oxic late also refers to growth under oxic conditions, but with cells at a D600 of 0.80. Low O2 refers to micro-oxic conditions.

Even though this nap is more highly expressed under denitrifying conditions, it still showed measurable expression under oxic conditions (Figure 1). Nitrite assays demonstrated that cells growing under oxic conditions in liquid medium have nitrate reductase activity; however, this activity was not dependent on nap, since similar levels of nitrite accumulated in a Nap-deficient mutant (Table 1). Increasing the ammonia level in the culture eliminated the nitrite accumulation under oxic conditions, indicating that this nitrite production under these conditions was due to Nas.

View this table:
Table 1 Nitrite production under oxic conditions by strains of A. tumefaciens and R. denitrificans

Nitrite production is indicated by +, with multiple + being used to indicate intensity. − indicates no nitrite production. All samples were taken between 12 and 24 h after inoculation. Potassium nitrate was added to 10 mM in all samples. S, succinate-based minimal medium (see [20]); S/N, succinate-based minimal medium with 3× ammonium; M, Difco Marine broth; wt, wild-type; nap−, nap-deficient strain.

Nitrite production was also observed in cells growing on plates under oxic conditions (Table 1). This activity was mainly due to Nap, since the nap mutant did not produce nitrite under these conditions unless the cells were incubated for >48 h. The wild-type strain produced nitrite within 24 h when grown on solid medium. A strain lacking nirK, the gene encoding Nir, produced the same amount of nitrite on plates as wild-type, indicating that the cells in the colony were not experiencing denitrifying conditions. The production of nitrite under these conditions is significant because it reveals that electron flow through Nap does not require reduced electron sources, but can occur with carbon sources that are the same oxidation state as the cell.

Nir and Nor regulation in select α-proteobacteria

Although the expression of nap has been shown to be variable, previous work with model organisms has consistently shown that the expression of the genes encoding Nir and Nor only occurs under micro-oxic conditions [16]. Therefore this section will focus mainly on regulatory mechanisms. Nir and Nor expression in R. sphaeroides 2.4.3 requires low oxygen and also requires the presence of a nitrogen oxide [17,18]. A nitrogen oxide is required because NO is a key effector. NO is probably detected by NnrR, a member of the FNR (fumarate nitrate reductase regulator)/CRP (cAMP receptor protein) family of transcriptional regulators (Figure 2). PrrBA is also required for expression of both Nir and Nor [19]. Evidence indicates that PrrBA directly regulates the gene encoding Nir, known as nirK, but not nor. However, nor expression still depends on PrrBA, since Nir generates the NO required for NnrR-dependent expression of nor.

Figure 2 Schematic of the regulation of nap, nirK and nor in R. sphaeroides

The nap regulatory scheme is taken from work with R. sphaeroides 2.4.1. nirK and nor regulation is taken from work with R. sphaeroides 2.4.3. Broken lines indicate only indirect evidence for the proposed interaction. The arrow at the bottom is used to distinguish the regulatory mechanisms in the proposed scheme.

A similar pattern of nirK and nor expression occurs in A. tumefaciens. This bacterium uses an orthologue of NnrR for nitrogen-dependent regulation [20]. An orthologue of the PrrBA system is also required for nirK expression [21]. In A. tumefaciens, this system has been designated ActRS. Purified ActR-P will bind to the nirK promoter, but not the nor promoter. This demonstrates that, as in R. sphaeroides, ActRS directly regulates nirK, but only indirectly regulates nor expression.

In these two denitrifiers, nirK and nor expression depends on at least two regulators, NnrR and PrrBA orthologues (Figure 2). The NnrR regulators serve as nitrogen oxide sensors. PrrBA and its orthologues appear to serve as sensors of the availability of terminal electron acceptors. Work in R. sphaeroides suggests that PrrBA activity is influenced by the activity of the cbb3 terminal oxidase [22,23]. This is a cytochrome c-dependent enzyme that reduces oxygen to water. Orthologues of this enzyme have been shown to have high affinity for oxygen [24]. When this enzyme is functioning, the PrrA/PrrA-P ratio is high. However, when oxygen is low, turnover of the enzyme is reduced, which leads to a decrease in the PrrA/PrrA-P ratio. In the related bacterium R. capsulatus, its PrrBA orthologues, designated RegBA, appear to be regulated by the redox status of the quinone pool [25]. Taken together, these results suggest that members of this family of regulators are influenced by electron flow through the respiratory chain. Therefore they act as sensors of the availability of electron acceptors, particularly oxygen. If oxygen, the preferred terminal electron acceptor, is present nirK is not expressed. However, these regulators do not appear to directly sense oxygen.

An additional layer of oxygen-dependent regulation?

Inactivation of genes encoding the cbb3 oxidase in R. sphaeroides has been shown to lead to high levels of PrrA-P under both oxic and micro-oxic conditions [26]. Loss of cbb3 oxidase activity would not be predicted then to decrease nirK expression in 2.4.3. Unexpectedly, loss of the cbb3 oxidase resulted in a significant decrease in nirK expression and Nir activity [19]. Recent results indicate that this phenotype is conditional [27]. In the original experiments, cells were grown under what is referred to as transition conditions in which the growth vessel is sealed with a rubber stopper after inoculation. Oxygen levels in the flask start out at atmospheric levels, but are steadily reduced as the cells grow. These conditions give rise to the Nir-deficient phenotype of the cbb3 mutant. However, when oxygen levels are decreased using gas mixes, the cbb3 mutant retains Nir activity at very low oxygen concentrations. The simplest explanation for this result is that, under transition conditions, loss of the cbb3 oxidase results in high residual oxygen levels in the culture, which prevents expression of nirK and nor. This oxygen-dependent inhibition is not observed using the low oxygen gas mixes. These results suggest that some aspect of nirK and nor expression is inhibited by oxygen. Nap expression, which is indirectly required as it is needed to produce the nitrite necessary for NO production, is not inhibited since nitrite accumulates in the cbb3 mutant [19]. PrrBA is not likely to be an oxygen-sensitive regulator since it is in its active form in this mutant. As NnrR is the only other known regulator, it may be oxygen-sensitive as previous work with an orthologue of NnrR has suggested [28] (Figure 2). There is also the possibility that an unidentified oxygen-sensitive protein is required for nirK and nor expression.

Gene expression and enzyme activity during transition from oxic to micro-oxic conditions

The expression and activity of Nir and Nor as a function of oxygen has been determined in A. tumefaciens [29]. Nitrate reductase expression and activity were not followed in these experiments, but the nap expression and activity results discussed above suggest that Nap will be present throughout the experiments (Figure 1). It seems unlikely, however, that Nap will be active under aqueous conditions without a significant reduction in oxygen levels. Once oxygen becomes limiting, this will allow diversion of electrons to Nap. Concomitant with this decrease in oxygen, the expression of nap will increase, allowing more electron flux through this branch of the respiratory pathway. nirK and nor are expressed once oxygen reaches <0.5 μM or approx. 0.2% of saturating oxygen. These oxygen concentrations are low enough to allow formation of ActR-P necessary for nirK activation. nirK expression is also dependent on NnrR which requires NO. However, since Nir is the primary source of NO, there must be some limited expression of nirK before oxygen levels reach sub-micromolar levels to produce the catalytic amounts of Nir necessary to potentiate the Nir–NO–NnrR-feedback loop. Consistent with this conclusion is the observation that transcripts of nirK are present before NO accumulation [29]. The expression of nor genes occurs subsequent to a burst of NO production. The onset of Nor translation brings the NO levels down to nanomolar levels consistent with balanced denitrification. Unexpectedly, the nor transcript was only present for a relatively short amount of time. It decreasesd to below detectable levels while NO was present in the headspace. nirK transcripts remained fairly steady throughout the anoxic portions of the experiments. The regulatory rationale for the relative instability of the norB transcript is not obvious.

Preliminary gas-phase measurements with R. sphaeroides 2.4.3 show a similar pattern (L. Bergaust and J.R. Shapleigh, unpublished work). In particular, there is a burst of NO that appears to be coincident with the onset of vigorous nitrite respiration. As observed previously with A. tumefaciens, transition between aerobic and anaerobic respiration was difficult under certain conditions [29]. These results suggest that, although some of the details may differ, expression of nap, nir and nor in R. sphaeroides 2.4.3 and A. tumefaciens follow similar patterns.

Aerobic denitrifiers

Work with most model denitrifiers is not supportive of the concept of aerobic denitrification, given the low oxygen levels required for expression of the genes encoding Nir and Nor. Nevertheless, studies finding evidence of aerobic denitrification are not uncommon [30,31]. As discussed above, there is no obvious a priori reason these studies should be dismissed. Moreover, there are many well known examples of organisms developing strategies to express seemingly anaerobic processes under aerobic conditions. A good example of this is the group of photosynthetic bacteria known as the aerobic anoxygenic phototrophs [32]. These bacteria express photosynthetic genes under oxic conditions, whereas most other related phototrophs are photosynthetic only under micro-oxic conditions. An example of this group is the bacterium R. denitrificans. As its name suggests, this bacterium is also a denitrifier and there are several studies that indicate this bacterium can express denitrification genes under oxic conditions [33,34].

In an effort to directly compare nitrogen oxide reductase activity in R. denitrificans with R. sphaeroides and A. tumefaciens, the former was grown under the same conditions used to assess nitrogen oxide activity in the latter. Under these conditions, it was observed that R. denitrificans produced nitrite during growth in either liquid or solid medium (Table 1). This bacterium has a membrane-bound Nar-type nitrate reductase, but lacks Nas. There was no evidence of nitrite reduction under oxic conditions in the dark. However, Nir was expressed in cells grown under these conditions, as demonstrated using a Methyl Viologen-dithionite-based Nir assay (results not shown). Nitrite was consumed by cells growing under photosynthetic conditions, consistent with previous results [34]. Attempts to grow R. denitrifcans under strictly denitrifying conditions failed. These results indicate that, under certain conditions, R. denitrificans might be an aerobic denitrifier. The exact physiological role of denitrification under these conditions is uncertain. It seems unlikely that nitrate respiration is being used for energetic purposes. Given the importance of redox homoeostasis in this group of proteobacteria, it is possible that denitrification is being used for redox balancing or some other accessory function.


Nap expression in the α-proteobacteria discussed in the present review seems to be adjusted to fit the physiological role it plays. In organisms where its primary function is redox homoeostasis it is expressed under oxic conditions, whereas if it is being used for respiration it is maximally expressed under denitrifying conditions. The expression of Nir and Nor is more tightly controlled through the action of at least two regulatory systems that limit expression to micro-oxic conditions. One regulator apparently senses NO, whereas the other is a more general regulator sensing availability of oxidant. Neither senses oxygen directly, but there may be an as yet undefined layer of oxygen-sensitive regulation. R. denitrificans may be an aerobic denitrifier, but the physiological role of the enzymes in this organism is uncertain.


This work was supported by the U.S. Department of Energy [grant number 95ER20206].


I acknowledge the work of Angela Hartsock and Armanda Roco.


  • Enzymology and Ecology of the Nitrogen Cycle: A Biochemical Society Focused Meeting held at University of Birmingham, U.K., 15–17 September 2010. Organized and Edited by Jeff Cole (University of Birmingham, U.K.), Rosa María Martínez-Espinosa (University of Alicante, Spain), David Richardson (University of East Anglia, Norwich, U.K.) and Nick Watmough (University of East Anglia, Norwich, U.K.).

Abbreviations: Nir, nitrite reductase; Nor, nitric oxide reductase; Nos, nitrous oxide reductase


View Abstract