Biochemical properties of Paracoccus denitrificans FnrP: reactions with molecular oxygen and nitric oxide
In Paracoccus denitrificans, three CRP/FNR family regulatory proteins, NarR, NnrR and FnrP, control the switch between aerobic and anaerobic (denitrification) respiration. FnrP is a [4Fe–4S] cluster-containing homologue of the archetypal O2 sensor FNR from E. coli and accordingly regulates genes encoding aerobic and anaerobic respiratory enzymes in response to O2, and also NO, availability. Here we show that FnrP undergoes O2-driven [4Fe–4S] to [2Fe–2S] cluster conversion that involves up to 2 O2 per cluster, with significant oxidation of released cluster sulfide to sulfane observed at higher O2 concentrations. The rate of the cluster reaction was found to be ~sixfold lower than that of E. coli FNR, suggesting that FnrP can remain transcriptionally active under microaerobic conditions. This is consistent with a role for FnrP in activating expression of the high O2 affinity cytochrome c oxidase under microaerobic conditions. Cluster conversion resulted in dissociation of the transcriptionally active FnrP dimer into monomers. Therefore, along with E. coli FNR, FnrP belongs to the subset of FNR proteins in which cluster type is correlated with association state. Interestingly, two key charged residues, Arg140 and Asp154, that have been shown to play key roles in the monomer–dimer equilibrium in E. coli FNR are not conserved in FnrP, indicating that different protomer interactions are important for this equilibrium. Finally, the FnrP [4Fe–4S] cluster is shown to undergo reaction with multiple NO molecules, resulting in iron nitrosyl species and dissociation into monomers.
KeywordsFumarate–nitrate reduction regulator Gene regulation Iron–sulfur cluster Oxygen Nitric oxide
Fumarate nitrate reduction
Liquid chromatography mass spectrometry
1-(Hydroxy-NNO-azoxy)-l-proline, disodium salt
Paracoccus denitrificans is a popular model organism and one of the best studied prokaryotes with respect to respiration. It has a remarkable metabolic versatility allowing it to thrive in aerobic or anaerobic environments [1, 2]. Under anaerobic conditions nitrogen oxides can be utilized as terminal electron acceptors, in place of oxygen, and P. denitrificans is one of several organisms for which the denitrification pathway is well understood. It expresses four essential reductases which sequentially reduce nitrate (nar), nitrite (nir), nitric oxide (nor) and nitrous oxide (nos) to dinitrogen . Optimal switching from aerobic respiration to the denitrification pathway is thus a key requirement for this flexibility, and the coordination of the denitrification enzymes is tightly controlled at the transcriptional level .
In E. coli, the fumarate and nitrate reduction (FNR) transcriptional regulator is responsible for sensing environmental levels of O2 and controlling the switch to anaerobic nitrate respiration . FNR proteins represent a major sub-group of the cyclic-AMP receptor protein (CRP) family of bacterial transcriptional regulators, and, like CRP, consist of two distinct domains that provide DNA-binding and sensory functions [6, 7, 8]. However, unlike E. coli, some bacterial species possess multiple members of the FNR protein family [4, 9]; P. denitrificans has three major FNR paralogues which coordinate the regulation of the denitrification enzymes . One of these paralogues, NarR, is a nitrate sensor involved in regulating nitrate reductase (nar) expression. The second, NnrR, is a heme-based nitric oxide sensor involved in the control of expression of nitrite (nir), nitric oxide (nor) and nitrous oxide reductases (nos) . The third, FnrP, is a true orthologue of E. coli FNR as it regulates genes encoding aerobic and anaerobic respiratory enzymes in response to O2 availability [4, 10, 11, 12].
Many bacteria that carry out anaerobic respiration with nitrate/nitrite as a terminal electron acceptor can produce nitric oxide (NO) endogenously [16, 17]. In P. denitrificans, NO production represents a key intermediary step during reduction of nitrate to dinitrogen . NO is a gaseous lipophilic radical that can function as both a signaling molecule and a cytotoxic agent . The latter arises from the reactivity of NO with a variety of important cellular targets [19, 20, 21], including iron–sulfur cluster proteins . Several regulatory proteins known to respond to NO contain iron–sulfur clusters as the sensory module [5, 23], and significant progress has been made recently in understanding the reactions of iron–sulfur clusters with NO in regulatory proteins [24, 25]. In the case of FnrP, there is growing in vivo evidence to suggest that, in addition to its primary function as an O2 sensor, it also plays a role in modulating gene expression in response to NO  in a similar way to E. coli FNR [25, 26].
Here we report investigations of the biochemical properties of FnrP. We present information on the nature of the iron–sulfur cluster, its reaction with O2 and NO, as well as the effect of these reactions on its association state. We compare our findings to those reported for E. coli FNR as well other FNR orthologues.
Materials and methods
Purification of FnrP
A GST-FnrP fusion protein (Fig. S1) was overproduced in aerobically grown E. coli BL21λDE3 harboring pSAD105, induced by the addition of 1 mM IPTG at 37 °C, as previously described . For the in vivo assembly of [4Fe–4S] FnrP, the aeration of cultures harboring pSAD105 was reduced after induction, as previously described  to mimic semi-aerobic conditions . FnrP was purified under anaerobic conditions using assay buffer (25 mM HEPES, 2.5 mM CaCl2, 100 mM NaCl, 100 mM NaNO3, pH 7.5) as previously described for E. coli FNR . FnrP was cleaved from the fusion protein using thrombin (Fig. S1), and, where necessary, the [4Fe–4S] cluster reconstituted, in vitro, as previously described [14, 27, 29], except that a 1 ml Q Sepharose column was used to concentrate the protein and assay buffer containing 500 mM KCl was used to elute the protein. Protein concentration was determined using the method of Bradford (BioRad), with bovine serum albumin as the standard . FnrP iron and acid-labile sulfide content were determined as previously described [31, 32].
UV–visible absorbance measurements were made with a Jasco V550 spectrometer. The extinction coefficient for the E. coli [4Fe–4S] FNR (ε 406 nm = 16,200 M−1 cm−1 ) was used to calculate the amount of [4Fe–4S] cluster present in FnrP samples. CD spectra were measured with a Jasco J810 spectropolarimeter. For liquid chromatography–mass spectrometry (LC–MS) an aliquot of FnrP (100 μL, 46 μM [4Fe–4S]) was combined with varying aliquots of aerobic (229 μM O2, 20 °C) or anaerobic assay buffer (200 μl final volume), and allowed to react for 15 min. Samples were diluted to ~2 μM final concentration, with an aqueous mixture of 1 % (v/v) acetonitrile, 0.3 % (v/v) formic acid, sealed, removed from the anaerobic cabinet and analyzed by an LC–MS instrument consisting of an Ultimate 3000 UHLPC system (Dionex, Leeds, UK), a ProSwift RP-1S column (4.6 × 50 mm) (Thermo Scientific), and a Bruker microQTOF-QIII mass spectrometer, running Hystar (Bruker Daltonics, Coventry, UK), as previously described .
FnrP samples (loaded at ~28 μM [4Fe–4S]) before and after exposure to O2 were analyzed by gel filtration under anaerobic conditions using assay buffer and a calibrated Sephacryl S-100HR 16/50 column (GE Healthcare), at a flow rate of 1 ml min−1.
Kinetic measurements were performed under pseudo-first-order conditions (162 μM O2) at 25 °C by combining varying ratios of aerobic and anaerobic assay buffer (2 ml total volume) with FnrP (8.5 μM [4Fe–4S]). Changes in A406 nm were used to track cluster conversion. A single or double exponential function, as necessary, was fitted to the data, as previously described [9, 28, 33]. Observed rate constants (k obs) obtained from the fits (in the case of double exponential fits, the rate constant for the first reaction phase was used) were divided by the O2 concentration, providing an estimate of the apparent second-order rate constant. Kinetic data fitting was performed using Origin (version 8, Origin Labs). Estimates of errors for rate constants are represented as ±the standard deviation.
Other analytical methods
FnrP (2 ml) was titrated against varying aliquots of O2 (~220 μM dissolved in assay buffer) or NO (as NONOate; Cayman chemicals), using anaerobic cuvettes and gas tight syringes. Stock solutions of the NO donor PROLI-NONOate (t½ = 1.5 s) were prepared in 50 mM NaOH and quantified optically (ε 252 nm 8400 M−1 cm−1).
Results and discussion
In vivo purified and in vitro reconstituted FnrP binds an identical [4Fe–4S] cluster
As FnrP contains seven cysteine residues (Cys8, 14, 17, 25, 28, 113 and 144), it is possible that differing methods (in vitro versus in vivo) of iron–sulfur cluster assembly might lead to significant differences in the ligation pattern and hence the local environment of the cluster. Since iron–sulfur clusters derive their optical activity from the fold of the protein to which they are ligated, the CD spectrum provides information about the cluster environment. The anaerobic CD spectra of both native and reconstituted [4Fe–4S] FnrP displays positive (+) features at 335 and 440 nm, together with negative (−) features at 300, 375, and 520 nm (see Fig. 2b). The Δε values for the native and reconstituted forms of FnrP are very similar, indicating that the [4Fe–4S] clusters are in essentially identical environments. Hence, reconstituted FnrP was used in subsequent experiments.
Interestingly, the shape of the CD spectrum of FnrP is quite distinct from that of E. coli FNR . The major bands at (−)375 nm and (+)440 nm, in FnrP, which originate primarily from S → Fe charge transitions, are equivalent to bands observed at (+)380 nm and (+)420 nm for [4Fe–4S] FNR and other [4Fe–4S] containing proteins, such as HiPIP, WhiD and ANR [9, 34, 35]. However, there is no strict correlation between the cluster type and the shape or sign of bands in the CD spectrum; presumably, differences in the cluster binding cavity and/or the geometry of the cluster lead to variation in the CD spectrum.
Reaction of [4Fe–4S] FnrP with O2 resembles that of E. coli FNR
FnrP reacts more slowly than E. coli FNR with O2 in vitro
Kinetic data for the reaction of FNR homologues with O2
Under O2-limiting conditions many bacteria induce high-affinity oxidases, as the ATP yield from oxygen respiration is significantly higher than anaerobic respiration . P. denitrificans is no different in this respect, with FnrP activating the expression of the cbb 3-type cytochrome c oxidase (cco) in vivo [2, 11]. We note that cco promoter activity increases by eight times during the switch from aerobic to semi-aerobic growth conditions . In contrast, the high-affinity E. coli cytochrome bd-I (cydAB) oxidase (a quinol:O2 oxidoreductase) is repressed by [4Fe–4S] FNR under anaerobic conditions [44, 45]. Thus, P. denitrificans may begin to utilize its high-affinity oxidases at significantly higher environmental O2 concentrations than its rivals to gain a competitive advantage.
The oligomeric state of FnrP is dependent on the [4Fe–4S] cluster
Moore and Kiley  showed that subunit interactions in E. coli FNR arise from a predominantly hydrophobic interface (see Fig. 6b), and that the negatively charged side chain of Asp154 is oriented towards this interface, where inter-subunit charge repulsion inhibits dimerization before cluster acquisition. Insertion of the [4Fe–4S]2+ cluster apparently causes shielding of the negative charge, by Ile-151, thereby facilitating dimerization. The recent structure of the E. coli FNR-like FNR from Aliivibrio fischeri  was broadly in agreement with this model, but indicated that hydrophobic contacts made between Ile-151 residues of the two subunits, rather than screening electrostatic repulsion, is key to stabilizing the dimer. Removal of the negatively charged side chain by substitution, such as in FNR-D154A, alleviates the repulsion even in the absence of a cluster, leading to a predominantly dimeric form whether the cluster is present or not. The positive charge of Arg140 is also important for FNR function. Substitution of this positively charged side chain, such as in FNR-R140A, resulted in an FNR variant with little anaerobic activity, implying a defect in dimerization. In the A. fischeri FNR structure, Arg-140 is located near the N terminus of the dimerization helix where no dimerization helix interactions occur. Instead, the Arg-140 sidechain forms a salt bridge with Asp-130 of the other subunit (B helix) and it was suggested that this interaction plays a key role in dimer stability. Interestingly, the double mutant FNR-R140A/D154A regained ≥70 % activity under anaerobic conditions . Assuming residues 130 to 149 participate in subunit interactions, analogous to E. coli FNR residues 140 to 159 (Fig. 6a), then Leu134, 141 and 148 presumably form the core of the hydrophobic interface in FnrP (see Fig. 6b). Moreover, charged residues Arg140 and Asp154, found to be important to the monomer/dimer equilibrium in E. coli FNR, appear to be replaced by Ala130 and Ala144 in FnrP. Thus, the ability of FnrP to undergo a cluster-induced monomer/dimer transition may depend on the lack of strong electrostatic interactions involving the side chains of Ala130 and Ala144 (cf the E. coli FNR variant R140A/D154A ). The remainder of the residues in FnrP appears to preserve the general nature of the E. coli FNR dimerization helix (see Fig. 6). We note that Bradyrhizobium japonicum FixK2 also contains Ala residues at equivalent positions to FnrP Ala130 and 144, and that other CRP/FNR family paralogues tend to contain an Ala residue in place of the Arg residue and a hydrophobic/non-charged residue in place of the negatively charged Asp residue (see Fig. S2).
FnrP is a nitric oxide (NO) sensor
Analysis by gel filtration of molecular mass changes upon nitrosylation of [4Fe–4S] FnrP revealed a decrease in mass from 51 to 33 kDa (Fig. 8d) for the bulk of the sample. A broad shoulder, corresponding to a mass of ~50–100 kDa, was also observed. The shoulder decreased following treatment with DTT, implying that a small proportion of the sample was in the form of a disulfide-bonded FnrP aggregate. Moreover, LC–MS analysis of nitrosylated FnrP revealed the presence of sulfur adducts (not shown), as previously observed for E. coli FNR . We conclude that nitrosylation of FnrP is likely to proceed in a manner similar to that previously reported for E. coli FNR, and that, like its reaction with O2, nitrosylation of FnrP results in dissociation of the dimer into monomers, again similar to E. coli FNR.
FNR proteins are global transcription factors that respond to changes in environmental O2 through the assembly and disassembly of an O2-sensitive [4Fe–4S] cluster. In the archetypal FNR protein E. coli FNR, molecular O2 brings about conversion of the [4Fe–4S] cluster into a [2Fe–2S] form, thereby triggering conformational changes that initiate monomerization and concomitant loss of sequence-specific DNA binding. In this respect, the P. denitrificans regulator FnrP is similar to its E. coli counterpart . However, sequence alignment revealed that the residues important for the monomer–dimer equilibrium for FnrP are different to those in E. coli FNR and involve fewer charged side chain interactions. Furthermore, we have found that the FnrP cluster is at least 6 times less sensitive to O2 than E. coli FNR. This finding is consistent with the observation of FnrP-activated expression of cbb 3-type cytochrome c oxidase, nitrate reductase and nitrous oxide reductase under semi-aerobic conditions in vivo [11, 48].
Many transcriptional regulators are known to respond to NO and in P. denitrificans the principal regulators are NsrR, NnrR, and FnrP. In other bacterial species NsrR regulates, amongst others, genes involved in detoxification, such as hmp, for which NO is a substrate [49, 50]. NnrR principally activates the expression of the nitrite, nitric oxide, and nitrous oxide reductases under anaerobic conditions in response to NO, thereby facilitating denitrification [13, 51, 52]. With regard to denitrification, FnrP co-regulates the expression of the nitrate and nitrous oxide reductases, ensuring their production under semi-aerobic conditions. We note that nitrate reductase is the most important source of endogenously derived NO during nitrate/nitrite respiration . It is suggested that if the NO detoxification systems are overwhelmed, FnrP will become nitrosylated leading to lowered expression of cco, and modulation of the nar and nos operons (that require [4Fe–4S] FnrP for activation). The concomitant detection of NO by NnrR would then ensure the timely expression of the nir, nor and nos operons, minimizing the transitory nitrosative stress as metabolic modes are switched over in favor of denitrification. Fig. S3 shows a summary of these regulatory systems. Here we have shown that [4Fe–4S] FnrP undergoes a nitrosylation reaction involving multiple NO molecules. This leads to dissociation of FnrP, containing iron–nitrosyl products similar to those observed for other NO-sensing iron–sulfur regulatory proteins, into monomers, providing a mechanistic basis for NO regulation of FnrP.
This work was supported by the UK’s Biotechnology and Biological Sciences Research Council grant BB/L007673/1 to NLB, AJT and JCC. We thank Nick Cull for technical assistance, Dr. Myles Cheesman for access to instrumentation and Prof. Stephen Spiro for pSAD105 encoding GST-FnrP.
Compliance with ethical standards
The authors dedicate this article to the memory of Bob Williams. They have been strongly influenced by his scientific work and inspired by his spirit. Andrew Thomson and Geoff Moore (see accompanying article on ferritins) studied in Oxford with Bob for their doctoral degrees. At UEA, together with the late Colin Greenwood, they established a multi-disciplinary research unit, the Centre for Metalloprotein Spectroscopy and Biology (CMSB). Nick Le Brun and Jason Crack carried out their doctoral studies in the CMSB under the supervision of Andrew and Geoff. Bob was an adviser to, and a strong supporter of, the CMSB.
- 12.Van Spanning RJ, De Boer AP, Reijnders WN, Westerhoff HV, Stouthamer AH, Van Der Oost J (1997) FnrP and NNR of Paracoccus denitrificans are both members of the FNR family of transcriptional activators but have distinct roles in respiratory adaptation in response to oxygen limitation. Mol Microbiol 23:893–907CrossRefPubMedGoogle Scholar
- 14.Hutchings MI, Crack JC, Shearer N, Thompson BJ, Thomson AJ, Spiro S (2002) Transcription factor FnrP from Paracoccus denitrificans contains an iron-sulfur cluster and is activated by anoxia: identification of essential cysteine residues. J Bacteriol 184:503–508PubMedCentralCrossRefPubMedGoogle Scholar
- 26.Pullan ST, Gidley MD, Jones RA, Barrett J, Stevanin TM, Read RC, Green J, Poole RK (2007) Nitric oxide in chemostat-cultured Escherichia coli is sensed by Fnr and other global regulators: unaltered methionine biosynthesis indicates lack of S nitrosation. J Bacteriol 189:1845–1855PubMedCentralCrossRefPubMedGoogle Scholar
- 41.Jervis AJ, Crack JC, White G, Artymiuk PJ, Cheesman MR, Thomson AJ, Le Brun NE, Green J (2009) The O2 sensitivity of the transcription factor FNR is controlled by Ser24 modulating the kinetics of [4Fe–4S] to [2Fe–2S] conversion. Proc Nat Acad Sci USA 106:4659–4664PubMedCentralCrossRefPubMedGoogle Scholar
- 49.Crack JC, Munnoch J, Dodd EL, Knowles F, Al Bassam MM, Kamali S, Holland AA, Cramer SP, Hamilton CJ, Johnson MK, Thomson AJ, Hutchings MI, Le Brun NE (2015) NsrR from Streptomyces coelicolor is a nitric oxide-sensing [4Fe–4S] cluster protein with a specialized regulatory function. J Biol Chem 290:12689–12704PubMedCentralCrossRefPubMedGoogle Scholar
- 52.van Spanning RJ, De Boer AP, Reijnders WN, Spiro S, Westerhoff HV, Stouthamer AH, van der Oost J (1995) Nitrite and nitric oxide reduction in Paracoccus denitrificans is under the control of NNR, a regulatory protein that belongs to the FNR family of transcriptional activators. FEBS Lett 360:151–154CrossRefPubMedGoogle Scholar
- 54.Nicholas KB, Nicholas HBJ (1997) Genedoc: a tool for editing and annotating multiple sequence alignments (Distributed by the authors) Google Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.