Reactivity of the N-terminal cysteine residues in active and inactive forms of FNR, and O2-responsive, Fe containing transcriptional regulator of Escherichia coli

Transcriptional Regulation of Gene Expression by Metalloproteins

FNR and SoxR are transcriptional regulators containing an iron–sulfur cluster. The iron–sulfur cluster in FNR acts as an oxygen sensor by reacting with oxygen. The structural change of the iron–sulfur cluster takes place when FNR senses oxygen, which regulates the transcriptional regulator activity of FNR through the change of the quaternary structure. SoxR contains the [2Fe–2S] cluster that regulates the transcriptional activator activity of SoxR. Only the oxidized SoxR containing the [2Fe–2S]2+ cluster is active as the transcriptional activator. CooA is a transcriptional activator containing a protoheme that acts as a CO sensor. CO is a physiological effector of CooA and regulates the transcriptional activator activity of CooA. In this review, the biochemical and biophysical properties of FNR, SoxR, and CooA are described.

In vivo cycling of the Escherichia coli transcription factor FNR between active and inactive states

FNR proteins are transcription regulators that sense changes in oxygen availability via assembly-disassembly of [4Fe-4S] clusters. The Escherichia coli FNR protein is present in bacteria grown under aerobic and anaerobic conditions. Under aerobic conditions, FNR is isolated as an inactive monomeric apoprotein, whereas under anaerobic conditions, FNR is present as an active dimeric holoprotein containing one [4Fe-4S] cluster per subunit. It has been suggested that the active and inactive forms of FNR are interconverted in vivo, or that iron-sulphur clusters are mostly incorporated into newly synthesized FNR. Here, experiments using a thermo-inducible fnr expression plasmid showed that a model FNR-dependent promoter is activated under anaerobic conditions by FNR that was synthesized under aerobic conditions. Immunoblots suggested that FNR was more prone to degradation under aerobic compared with anaerobic conditions, and that the ClpXP protease contributes to this degradation. Nevertheless, FNR was sufficiently long lived (half-life under aerobic conditions, approximately 45 min) to allow cycling between active and inactive forms. Measuring the abundance of the FNR-activated dms transcript when chloramphenicol-treated cultures were switched between aerobic and anaerobic conditions showed that it increased when cultures were switched to anaerobic conditions, and decreased when aerobic conditions were restored. In contrast, measurement of the abundance of the FNR-repressed ndh transcript under the same conditions showed that it decreased upon switching to anaerobic conditions, and then increased when aerobic conditions were restored. The abundance of the FNR- and oxygen-independent tatE transcript was unaffected by changes in oxygen availability. Thus, the simplest explanation for the observations reported here is that the FNR protein can be switched between inactive and active forms in vivo in the absence of de novo protein synthesis.

DNA target sequence and FNR-dependent gene expression

FNR proteins are global transcription regulators that respond to fluctuations in environmental oxygen. They recognise a DNA target consisting of an inverted repeat, TTGATN(1)N(2)N(3)N(4)ATCAA (where N(1-4) represents a non-conserved tetrad, NCT). Analysis of 68 known and predicted FNR sites from the Escherichia coli K12 genome revealed a bias toward A or T at positions N(2) and N(3) of the NCT. The effect of the NCT sequence on FNR-dependent transcription in vivo was assessed using a series of class II and class I model promoters with different NCT sequences. Changing the NCT sequence did not affect basal activity but altered anaerobic induction by as much as an order of magnitude. Thus, the NCT sequence is a fundamental component in setting the dynamic range of the FNR switch.

Miscoordination of the iron-sulfur clusters of the anaerobic transcription factor, FNR, allows simple repression but not activation

The FNR protein of Escherichia coli regulates target genes in response to anaerobiosis. Environmental oxygen is sensed by the acquisition of oxygen-labile [4Fe-4S] clusters that promote dimerization, DNA binding, and productive interactions with RNA polymerase. Three N-terminal cysteine residues (Cys(20), Cys(23), and Cys(29)) and Cys(122) act as ligands for the FNR iron sulfur clusters. An FNR variant, FNR-C20S, that retains only trace activity in vivo can acquire [4Fe-4S] clusters in vitro that enhance site-specific DNA binding. Second site substitutions in activating regions AR1, AR2, and AR3 restore in vivo activity to FNR-C20S, suggesting that the impairment in FNR-C20S activity is due to a failure to communicate with RNA polymerase effectively. Here we show that FNR-C20S can repress a simple FNR-regulated promoter in vivo and that it can form productive heterodimers with an FNR variant with altered DNA binding specificity, FNR-E209V. Transcription studies with FNR-E209V.FNR-C20S heterodimers indicate that the presence of a miscoordinated iron-sulfur cluster (FNR-C20S) in the downstream (but not the upstream) subunit of the FNR dimer impairs activation from a class II promoter and that this impairment can be overcome by amino acid substitutions known to unmask AR2 or improve AR3 in the affected subunit.

Anaerobic acquisition of [4FE 4S] clusters by the inactive FNR(C20S) variant and restoration of activity by second-site amino acid substitutions

The FNR protein of Escherichia coli controls the transcription of target genes in response to anoxia. The anaerobic incorporation of oxygen-sensitive [4Fe 4S] clusters promotes dimerization, which in turn enhances DNA binding. Four potential iron ligands (C20, C23, C29 and C122) are essential for normal FNR activity in vivo. Three FNR variants (C20S, C23G and C29G) retained the ability to incorporate oxygen-sensitive [4Fe 4S] clusters and to bind target DNA with essentially unimpaired affinity, suggesting that their failure to function normally in vivo resides at a later stage in the signal transduction pathway. The C122 variant failed to assemble iron-sulphur clusters and to bind DNA. Second-site substitutions that partially restore activity to FNR(C20S) were generated by error-prone polymerase chain reaction and were located in the dimer interface, in the activating regions (AR1, 2 or 3) or close to C122. Substitutions at E47, R48, E123, I124, E127 or T128 allowed the extent of the FNR AR2 surface to be defined. Only one revertant, FNR(C20S Y69F G149S), specifically corrected the C20S defect. It was concluded that [4Fe 4S] cluster acquisition, dimerization and DNA binding are not sufficient to confer transcription regulatory activity on FNR: the iron-sulphur cluster must also be correctly liganded in order to establish effective activating contacts between FNR and RNA polymerase.

Molecular genetics of the genus Paracoccus: metabolically versatile bacteria with bioenergetic flexibility

Paracoccus denitrificans and its near relative Paracoccus versutus (formerly known as Thiobacilllus versutus) have been attracting increasing attention because the aerobic respiratory system of P. denitrificans has long been regarded as a model for that of the mitochondrion, with which there are many components (e.g., cytochrome aa3 oxidase) in common. Members of the genus exhibit a great range of metabolic flexibility, particularly with respect to processes involving respiration. Prominent examples of flexibility are the use in denitrification of nitrate, nitrite, nitrous oxide, and nitric oxide as alternative electron acceptors to oxygen and the ability to use C1 compounds (e.g., methanol and methylamine) as electron donors to the respiratory chains. The proteins required for these respiratory processes are not constitutive, and the underlying complex regulatory systems that regulate their expression are beginning to be unraveled. There has been uncertainty about whether transcription in a member of the alpha-3 Proteobacteria such as P. denitrificans involves a conventional sigma70-type RNA polymerase, especially since canonical -35 and -10 DNA binding sites have not been readily identified. In this review, we argue that many genes, in particular those encoding constitutive proteins, may be under the control of a sigma70 RNA polymerase very closely related to that of Rhodobacter capsulatus. While the main focus is on the structure and regulation of genes coding for products involved in respiratory processes in Paracoccus, the current state of knowledge of the components of such respiratory pathways, and their biogenesis, is also reviewed.

Alternative respiratory pathways of Escherichia coli: energetics and transcriptional regulation in response to electron acceptors

The electron-transport chains of Escherichia coli are composed of many different dehydrogenases and terminal reductases (or oxidases) which are linked by quinones (ubiquinone, menaquinone and demethylmenaquinone). Quinol:cytochrome c oxido-reductase ('bc1 complex') is not present. For various electron acceptors (O2, nitrate) and donors (formate, H2, NADH, glycerol-3-P) isoenzymes are present. The enzymes show great variability in membrane topology and energy conservation. Energy is conserved by conformational proton pumps, or by arrangement of substrate sites on opposite sides of the membrane resulting in charge separation. Depending on the enzymes and isoenzymes used, the H+/e- ratios are between 0 and 4 H+/e- for the overall chain. The expression of the terminal reductases is regulated by electron acceptors. O2 is the preferred electron acceptor and represses the terminal reductases of anaerobic respiration. In anaerobic respiration, nitrate represses other terminal reductases, such as fumarate or DMSO reductases. Energy conservation is maximal with O2 and lowest with fumarate. By this regulation pathways with high ATP or growth yields are favoured. The expression of the dehydrogenases is regulated by the electron acceptors, too. In aerobic growth, non-coupling dehydrogenases are expressed and used preferentially, whereas in fumarate or DMSO respiration coupling dehydrogenases are essential. Coupling and non-coupling isoenzymes are expressed correspondingly. Thus the rationale for expression of the dehydrogenases is not maximal energy yield, but could be maximal flux or growth rates. Nitrate regulation is effected by two-component signal transfer systems with membraneous nitrate/nitrite sensors (NarX, NarQ) and cytoplasmic response regulators (NarL, NarP) which communicate by protein phosphorylation. O2 regulates by a two-component regulatory system consisting of a membraneous sensor (ArcB) and a response regulator (ArcA). ArcA is the major regulator of aerobic metabolism and represses the genes of aerobic metabolism under anaerobic conditions. FNR is a cytoplasmic O2 responsive regulator with a sensory and a regulatory DNA-binding domain. FNR is the regulator of genes required for anaerobic respiration and related pathways. The binding sites of NarL, NarP, ArcA and FNR are characterized for various promoters. Most of the genes are regulated by more than one of the regulators, which can act in any combination and in a positive or negative mode. By this the hierarchical expression of the genes in response to the electron acceptors is achieved. FNR is located in the cytoplasm and contains a 4Fe4S cluster in the sensory domain. The regulatory concentrations of O2 are 1-5 mbar. Under these conditions O2 diffuses to the cytoplasm and is able to react directly with FNR without involvement of other specific enzymes or protein mediators. By oxidation of the FeS cluster, FNR is converted to the inactive state in a reversible process. Reductive activation could be achieved by cellular reductants in the absence of O2. In addition, O2 may cause destruction and loss of the FeS cluster. It is not known whether this process is required for regulation of FNR function.

O2 as the regulatory signal for FNR-dependent gene regulation in Escherichia coli

With an oxystat, changes in the pattern of expression of FNR-dependent genes from Escherichia coli were studied as a function of the O2 tension (pO2) in the medium. Expression of all four tested genes was decreased by increasing O2. However, the pO2 values that gave rise to half-maximal repression (pO(0.5)) were dependent on the particular promoter and varied between 1 and 5 millibars (1 bar = 10(5) Pa). The pO(0.5) value for the ArcA-regulated succinate dehydrogenase genes was in the same range (pO(0.5) = 4.6 millibars). At these pO2 values, the cytoplasm can be calculated to be well supplied with O2 by diffusion. Therefore, intracellular O2 could provide the signal to FNR, suggesting that there is no need for a signal transfer chain. Genetic inactivation of the enzymes and coenzymes of aerobic respiration had no or limited effects on the pO(0.5) of FNR-regulated genes. Thus, neither the components of aerobic respiration nor their redox state are the primary sites for O2 sensing, supporting the significance of intracellular O2. Non-redox-active, structural O2 analogs like CO, CN-, and N3-, could not mimic the effect of O2 on FNR-regulated genes under anaerobic conditions and did not decrease the inhibitory effect of O2 under aerobic conditions.