The cd1 NiRs (nitrite reductases) are enzymes catalysing the reduction of nitrite to NO (nitric oxide) in the bacterial energy conversion denitrification process. These enzymes contain two distinct redox centres: one covalently bound c-haem, which is reduced by external electron donors, and another peculiar porphyrin, the d1-haem (3,8-dioxo-17-acrylate-porphyrindione), where nitrite is reduced to NO. In the present paper, we summarize the most recent results on the mechanism of nitrite reduction by the cd1 NiR from Pseudomonas aeruginosa. We discuss the essential catalytic features of this enzyme, with special attention to the allosteric regulation of the enzyme's activity and to the mechanism employed to avoid product inhibition, i.e. trapping of the active-site reduced haem by the product NO. These results shed light on the reactivity of cd1 NiRs and assign a central role to the unique d1-haem, present only in this class of enzymes.
Pseudomonas aeruginosa is a Gram-negative bacterium commonly found in soil and water, well known for its metabolic versatility; under anaerobic conditions it can use nitrate and nitrite to produce energy via the denitrification pathway. In natural environment, denitrification is the part of the biological nitrogen cycle in which nitrate is transformed into nitrogen gas; reduction of nitrate occurs in four stages each catalysed by a specific metalloenzyme [i.e. the nitrate, nitrite, NO (nitric oxide) and N2O reductases] , according to the following scheme: NO−3→NO−2→NO→N2O→N2.
The denitrification pathway is expressed under low oxygen tension in the presence of nitrogen oxides such as nitrate and nitrite .
It is well known that facultative aerobic bacteria are capable of adapting their respiratory capability and reorganizing their metabolism to cope with O2 availability. A striking example is given by denitrifiers, including human pathogens such as Ps. aeruginosa and Neisseria meningitidis, which were shown to adapt to the variable supply of O2 and nitrate/nitrite by regulating the expression and activity of the various reductases involved in respiratory metabolism.
In the denitrification pathway, NO is produced from nitrite by the enzyme NiR (nitrite reductase); two distinct classes of NiRs are found in bacteria both yielding NO as the main product. The two types of NiR contain either copper (CuNiR) or haem (cd1NiR) as cofactor, with the haem-containing enzymes occurring more frequently; in Ps. aeruginosa, nitrite reduction is carried out by cd1NiR. The membrane-bound enzyme NOR (nitric oxide reductase) is responsible for NO conversion into N2O.
Ps. aeruginosa is also a human pathogen, responsible for severe nosocomial infections, in particular in chronic respiratory diseases such as cystic fibrosis. In the airways of the patients, Ps. aeruginosa is capable of anaerobic growth by respiration using nitrate or nitrite as terminal electron acceptors ; thus pathogenesis, NO metabolism and anaerobic denitrification are strictly related [2,3]. Denitrification is important to detoxify NO under anaerobic conditions: the host's defence systems are neutralized by the activity of NOR . On the other hand, the denitrification machinery is a key signal for the expression of virulence factors; in a recent paper, van Alst et al.  demonstrated that production of NO by cd1NiR is required under aerobic conditions for Ps. aeruginosa virulence in a Caenorhabditis elegans model. Given that excess free NO may cause metabolic suicide of the pathogen , during anaerobic growth, Ps. aeruginosa controls the levels of NO by regulating its synthesis and degradation . The key role of nitrite reduction is also highlighted by experiments showing that exposure to 15 mM NO2− at pH 6.5 kills a mucoid variant of Ps. aeruginosa, abundant in the airways of cystic fibrosis patients, in part because of the low NiR activity present in this mutant strain [7,8].
In the present paper, we describe the mechanism of nitrite reduction by cd1NiR from Ps. aeruginosa (Pa-cd1NiR) [9,10]. We discuss the essential catalytic features of this enzyme, with special attention to the recent experiments on the allosteric regulation of the activity and to the mechanism employed to avoid product inhibition, i.e. trapping of the active-site reduced haem by the product NO.
Pa-cd1NiR is a homodimer containing one c-haem and one d1-haem (3,8-dioxo-17-acrylate-porphyrindione) group in each subunit. Electrons are transferred from cytochrome c551 to the c-haem moiety of the enzyme , and thereby internally to the active-site Fe(III) d1-haem, to which the substrate NO2− binds and is reduced to NO . The d1-haem (Figure 1A) is a partially saturated macrocycle with a set of oxo, methyl and acrylate substituents; other haems in biology in which the porphyrin ring is partially saturated are the d-haem in Escherichia coli and sirohaem of bacterial and plant sulfite and NiRs (Figures 1C and 1D) [12,13]. The d1-haem is unique to the cd1NiRs [1,14] and is synthesized by a specialized pathway present only in denitrifiers (strongly induced in Ps. aeruginosa upon nitrite treatment).
Chemical structure of the
d1-haem and other partially saturated porphyrins:
Nitrite reduction by
Nitrite reduction to NO is the physiologically relevant activity of cd1NiR [15,16]; indeed the expression of cd1NiR is induced by low O2 tensions and presence of nitrogen oxides . NO is produced efficiently by Pa-cd1NiR (catalytic-centre activity=6 s−1 at pH 7.0)  and the activity is pH dependent with an optimum between pH 5.8 and pH 6.5 [15,17]. The current knowledge on the individual steps in the catalytic cycle is summarized below.
In cd1NiR, nitrite binds to the reduced d1-haem with high affinity (Km=6 μM) ; binding of nitrite to the Fe(II)-iron is expected in the reaction mechanism since this state of the iron has to supply the electron needed for the reduction process. The nitrite molecule is thought to bind via the nitrogen atom in Pa-cd1NiR forming the so-called nitro complex; this is an important point to clarify since the chemical reactivity of the nitrite molecule may differ considerably depending on the mode of binding. The crystal structure of the reduced nitrite-bound derivative of Pa-cd1NiR is not available, but in the Paracoccus pantotrophus cd1NiR the nitrite molecule binds via the nitrogen atom . This observation agrees with results on other haem proteins [20,21] and on synthetic iron porphyrin nitrite complexes in which, regardless of the iron oxidation state, the N-binding mode is observed [22–24]. The nitrite N-binding mode also agrees with the current mechanism of the reduction of nitrite by cd1NiRs which is thought to occur via a double protonation of a terminal O atom of the nitrite molecule. Theoretical calculations have suggested that the O-binding mode is also possible for nitrite binding and reduction of cd1NiR  and also for other haem proteins such as haemoglobin . Although nitrite can bind through the O-binding mode (the so-called nitrito mode) to the haem of myoglobin and haemoglobin [27–29], in the case of the d1-haem there is no experimental evidence that such O-binding mode may occur.
The high affinity for nitrite of the ferrous d1-haem (see above) is a peculiar and physiologically relevant feature of all cd1NiRs; indeed these proteins display unusually high affinity for several anionic molecules in the ferrous state, including the inhibitor cyanide [30,31]. This behaviour is remarkably different from that observed in the b-type haem-containing proteins in which the negatively charged molecules (nitrite and cyanide) usually bind much better to the Fe(III) iron. The much higher affinity for nitrite of the d1-haem can be partially explained with the presence of two electron withdrawing carbonyl groups on the d1-haem ring (Figure 1A). The two conserved histidine residues (His327 and His369) in the active-site pocket  (see Figure 2) also contribute to the stabilization of the nitrite anion: upon mutagenesis of the latter histidine residues into alanine, the affinity for nitrite falls and the turnover rate is reduced 100-fold . The high affinity for nitrite of reduced Pa-cd1NiR is important for the catalytic mechanism as discussed below.
Conformational changes seen after the reduction of Pa-
Electron transfer from the
c-haem to the d1-haem
As mentioned above, in the catalytic cycle the substrate (i.e. nitrite) binds to the ferrous enzyme and is then reduced to yield NO and oxidized d1-haem. Reduction of the d1-haem occurs (by intramolecular electron transfer) from the c-haem, which in turn is reduced by external electron donors (see above). Early kinetic studies of cd1NiRs isolated from Ps. aeruginosa , P. pantotrophus  and Pseudomonas stutzeri  yielded considerably different rate constants for this step: 3, 1400 and 23 s−1 respectively. Interestingly, the internal electron transfer rate in Pa-cd1NiR is very similar to that of the overall turnover of the enzyme at pH 7.0 (6 s−1) , suggesting that this step may kinetically control the rate of reaction.
Recently, the reduction process of Pa-cd1NiR was investigated using the pulse radiolysis method by employing 1-MNA* (1-methylnicotinamide) and CO2− radicals as reductants ; the results provide clear kinetic evidence for negative co-operativity between the two d1-haem sites in the dimer and suggest that, in Pa-cd1NiR, the internal electron transfer rate from the c-haem to the d1-haem is tightly regulated by an allosteric mechanism. In more detail, the c-haem to d1-haem electron transfer rate constant was found to decrease (by more than one order of magnitude) as the number of electrons introduced in the enzyme is increased . This decrease in rate may, in principle, arise from a decrease in the electronic coupling between donor and acceptor sites due to conformational changes between reduced and oxidized Pa-cd1NiR . These conformational changes include the relocation of the side chain of Tyr10 located nearby the d1-haem, the coupled dissociation of the OH− ligand of the Fe(III) iron and the disruption of the hydrogen bonds at the interface between the two haem domains (Figure 2). In agreement with this structural interpretation, the intramolecular electron transfer is faster (approx. 10-fold) when the conserved His369 is replaced by alanine . The positive charge density on the distal side of the d1-haem is decreased in this mutant and thus the hydroxide ligand is destabilized, lowering the energy barrier for electron transfer and thus increasing the rate.
In summary, we propose that, in Pa-cd1NiR, allosteric communication between identical monomers may depend on a perturbation of the domain–domain interface within each monomer and a coupled large relocation of the c-haem domain  on reduction of the enzyme. These conformational changes disrupt the above-mentioned hydrogen bonds, decreasing the electronic coupling between the c-haem and the d1-haem. Thus the control of the electron transfer rates in Pa-cd1NiR involves contributions from local changes, affecting the driving force of the intramolecular electron transfer, and more global changes, which may affect the electronic coupling between sites (within and between monomers), further rising the energy barrier for electron transfer as the enzyme undergoes reduction.
The complex allosteric behaviour observed for the homodimeric Pa-cd1NiR indirectly confirms in our opinion the relevance of this step in the control of the reactivity of the enzyme.
Catalysis and product release
In the catalytic cycle, the formation of a complex between NO and the reduced d1-haem might slow down product release; indeed, the reduced d1-haem–NO complex may be considered an inhibited species  given that NO binds tightly to ferrous haem proteins [40,41]. However, we have clearly shown that the rate constant of NO dissociation from reduced Pa-cd1NiR is high (up to 70 s−1) , several orders of magnitude greater than that of haem b-containing proteins. Consequently, the affinity of Pa-cd1NiR for NO is lowered and the ferrous enzyme is not firmly inhibited by NO [17,42]. It is noteworthy that nitrite reduction can still be monitored after pre-incubation of reduced Pa-cd1NiR with NO .
These conclusions on Pa-cd1NiR are also supported by the experiments carried out on the enzyme from P. pantotrophus suggesting that intramolecular c-haem to d1-haem electron transfer triggers product release ; indeed it has been shown that P. pantotrophus d1Fe(III)–NO complex is a very long-lived species in the absence of excess reductant [44,45]. Therefore also P. pantotrophus cd1NiR works efficiently only in the presence of substrate and electron donor proteins [16,44,46], i.e. it cannot release NO in the absence of reducing equivalents.
We have also shown that nitrite can displace the NO bound to the ferrous enzyme , allowing the enzyme to enter a new catalytic cycle; therefore the high affinity of the active-site ferrous d1-haem for nitrite (see above) actively contributes to NO dissociation during the catalytic cycle. In agreement with this observation, if the affinity of Pa-cd1NiR for nitrite is decreased (as in the H369A mutant) the fully reduced NO-bound derivative accumulates [18,42]. The observation that NO and nitrite can compete for binding may suggest that the formation of N2O3 could in principle occur, for example, in a reaction similar to that proposed for Hb . This event is, however, highly unlikely, mainly because during catalysis the d1-haem is maintained in the reduced state by internal electron transfer from the c-haem. Moreover, ferric d1-haem has low affinity for nitrite , a feature that probably limits further reaction with free NO to produce N2O3.
In summary, the high affinity for nitrite, the allosteric control of the internal electron transfer rate and the exceptionally high NO dissociation rate ascertain that, when nitrite is available, the NiR is active and can actively produce and release NO.
In facultative aerobic bacteria, such as Ps. aeruginosa, nitrite is used as an electron acceptor and converted into NO by the specialized class of haem-containing enzymes called cd1NiRs . The mechanism employed by cd1NiRs is now becoming clear, highlighting the essential role of the peculiar d1-haem, whose presence in these enzymes was largely unexplained. As mentioned above, the high affinity of reduced d1-haem for nitrite  and the exceptionally fast NO dissociation  are trademarks of the d1-haem, and are tuned by two conserved histidine residues present in the active site of Pa-cd1NiR.
We previously proposed  that ancient haems, such as the d1-haem of cd1NiR or the sirohaem of bacterial and plant nitrite and sulfite reductases, have been evolutionaily conserved because their role is strategic to the organism where they are found today. The peculiar low affinity for NO of the reduced d1-haem of Ps. aeruginosa cd1NiR is a component of this strategy, which could not be achieved by the more common b-type haem. It would be very interesting to measure and compare the NO releasing properties of the various porphyrins shown in Figure 1; such experiments, together with a computational study of the electronic structure of the same porphyrins may allow us to correlate the functional properties of these partially saturated macrocycles with NO with their structure.
Finally, the recent evidence that nitrite can also be reduced to NO under hypoxic conditions in eukaryotes, suggests that anaerobic/microaerobic nitrite reduction in higher organisms is a vestigial function originating from the denitrifying metabolism and operative long before the advent of aerobic respiration .
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.).
This work was funded by the Ministero della Università of Italy [projects 20074TJ3ZB and RBRN07BMCT] and from the University of Rome La Sapienza.