NO sensing by FNR: regulation of the Escherichia coli NO-detoxifying flavohaemoglobin, Hmp

Abstract

Nitric oxide (NO) is a signalling and defence molecule of major importance in biology. The flavohaemoglobin Hmp of Escherichia coli is involved in protective responses to NO. Because hmp gene transcription is repressed by the O(2)-responsive regulator FNR, we investigated whether FNR also senses NO. The [4Fe-4S](2+) cluster of FNR is oxygen labile and controls protein dimerization and site-specific DNA binding. NO reacts anaerobically with the Fe-S cluster of purified FNR, generating spectral changes consistent with formation of a dinitrosyl-iron-cysteine complex. NO-inactivated FNR can be reconstituted, suggesting physiological relevance. FNR binds at an FNR box within the hmp promoter (P(hmp)). FNR samples inactivated by either O(2) or NO bind specifically to P(hmp), but with lower affinity. Dose-dependent up-regulation of P(hmp) in vivo by NO concentrations of pathophysiological relevance is abolished by fnr mutation, and NO also modulates expression from model FNR-regulated promoters. Thus, FNR can respond to not only O(2), but also NO, with major implications for global gene regulation in bacteria. We propose an NO-mediated mechanism of hmp regulation by which E.coli responds to NO challenge.

Fig. 1. Effect of NO on FNR. (A) Optical spectra of [4Fe–4S]²⁺ FNR (27 µM, monomer) before and after treatment with aliquots of 4.17 mM proline NONOate to give a final concentration of 140 µM NO (72 µM NONOate). Inset: a plot of ΔA₃₆₀ versus the ratio of [NO]/[FNR]. When no further change was apparent, a sample was withdrawn for EPR analysis. (B) Optical spectra of the same FNR suspension acquired after sequential additions of proline NONOate. The lowest spectrum is FNR in the absence of NO and the uppermost one is at a [NO]/[FNR] ratio of 5.2. The [NO]/[FNR] ratio increases from 0, 0.2, 0.3, 0.6, 1.2, 2.2, 2.9, 3.7 to 5.2, giving rise to dose-dependent absorbance increases at 360 nm.

Fig. 2. Optical spectra of DNIC complexes. (A) Model complexes. The monomeric species (445 µM; solid line) displays characteristic absorption maxima at 397 nm (ε ∼3580/M/cm), 603 nm (ε ∼299/M/cm) and 772 nm (ε ∼312/M/cm) (Costanzo et al., 2001). The dimeric species (250 µM, dashed line) displays characteristic absorption maxima at 310 nm (ε ∼9650/M/cm), 362 nm (ε ∼8529/M/cm) and 750 nm (ε ∼100 M/cm). Inset: structures proposed for monomeric and dimeric DNIC species (Costanzo et al., 2001). (B) Comparison between experimental (dashed line) and simulated (black line) spectra of DNICs. The simulated spectrum consists of contributions from 20 and 80% of the monomeric and dimeric DNIC species, respectively, plus apoFNR.

Fig. 3. EPR spectra of [4Fe–4S]²⁺ FNR. (A) The FNR sample used in Figure 1, after treatment with 140 µM NO (72 µM proline NONOate; 27 µM FNR monomer). (B) Increasing magnitudes of EPR signals elicited by addition of aliquots of NO solution (NO and FNR concentrations were as indicated). The temperature was 77 K. Microwave power and frequency were in (A) 2.007 mW and 9.653 GHz, and in (B) 2.000 mW and 9.669 GHz, respectively. Modulation amplitude, frequency and receiver gain were in (A) 5 gauss (0.5 mT), 100 kHz and 5.02 × 10⁵ and in (B) 10 gauss (1 mT), 100 kHz and 1.00 × 10⁵, respectively.

Fig. 4. DNA binding by holo-FNR, O₂- and NO-treated FNR proteins. Gel retardation assays were performed by incubating the ³²P-labelled P_hmp DNA fragment (wild-type) with increasing amounts of (A) holo-FNR, (B) O₂-treated FNR and (C) NO-treated FNR protein. The ³²P- labelled P_hmp DNA fragment (mutated) was incubated with increasing amounts of (D) holo-FNR. Arrows indicate the FNR–target DNA (black) and the FNR–P_hmp complex (white). Protein concentrations (monomer) are shown.

Fig. 5. DNase I protection by holo-FNR, NO-treated FNR and RNAP bound to P_hmp. The ³²P-labelled P_hmp DNA fragment was digested with DNase I in the presence of various concentrations of (A and C) holo-FNR, (B) NO-treated FNR and (C) RNAP. Template strand patterns shown were obtained with no FNR (lanes -), 0.2 µM (lanes 1), 2 µM (lanes 2), 12 µM (lanes 3) and 20 µM (lanes 4) FNR protein (as monomer), and with 69 nM RNAP (lane 5). Lanes G provide a calibration for the GC base pairs in the P_hmp region. Hypersensitive sites are marked with black or white arrowheads. Hollow boxes show the region protected by dimeric FNR binding. Striped boxes show the position of a DNase I-protected AT-rich hexamer. The filled boxes indicate the region bound by RNAP σ⁷⁰ subunit. The –10 and –35 hexamers, the ribosome-binding site (rbs), the transcription start site (+1) and the FNR box are indicated on the P_hmp sequence shown below.

Fig. 6. NO derepresses hmp expression in vivo by reacting with FNR. (A) Release of NO by the NOC mixture over 1 h. Traces 1 and 2 are the theoretical NO release patterns at 37°C (in PBS pH 7.4) by NOC-7 (half-life 5 min) and NOC-5 (half-life 25 min), respectively. Trace 3 is the total NO release by the NOC mixture (addition of traces 1 and 2). Trace 4 is the measured NO release by the NOC mixture under the same conditions of NO challenge in anoxic LB-containing cells (OD₆₀₀ of ∼0.2) at 37°C. Anaerobic cultures of (B) RKP2178 [Φ(hmp-lacZ)1] and (C) RKP2185 (RKP2178 but fnr-271::Tn10) were challenged with different amounts of NOC mixture. The β-galactosidase activity is the mean of three independent repetitions of the experiment; bars show standard deviations.

Fig. 7. NO modulates gene expression in vivo by reacting with FNR. Anaerobic cultures of E.coli strains carrying (A) the FNR-inducible promoter [FF(–71.5)::lacZ] or (B) the FNR-repressible promoter (FFgalΔ4::lacZ) were treated with different amounts of NO. β-galactosidase activity is the mean of three independent experiments.

Fig. 8. NO-mediated mechanism of anaerobic up-regulation of the E.coli hmp gene. FNR forms are represented by ovals. Cubes and hexagons represent the [4Fe–4S]²⁺ and nitrosylated clusters, respectively. DNA (P_hmp), NO and Hmp molecules are also shown. (A) Holo-FNR–DNA binding and repression of hmp under anaerobic conditions. (B) FNR senses NO by reaction with [4Fe–4S]²⁺ clusters to generate DNIC–FNR. This post-transcriptional modification is likely to promote monomerization of FNR. (C) NO-treated FNR (and apo-FNR) can eventually bind P_hmp with a lower affinity to avoid sudden and complete derepression of the hmp gene. (D) The NO signal transduction through FNR allows derepression of hmp transcription. Synthesized Hmp protein detoxifies NO to produce N₂O or NO₃^– in the absence or presence of O₂, respectively (Poole and Hughes, 2000). (E) When the NO concentration is reduced, putative Fe–S cluster-repairing mechanisms can reconstitute the NO sensor, FNR.