OUP user menu

Regulation of the flavorubredoxin nitric oxide reductase gene in Escherichia coli: nitrate repression, nitrite induction, and possible post-transcription control

Patrícia N. da Costa, Miguel Teixeira, Lígia M. Saraiva
DOI: http://dx.doi.org/10.1016/S0378-1097(02)01186-2 385-393 First published online: 1 January 2003


Escherichia coli flavorubredoxin is a new type of cytoplasmic nitric oxide (NO) reductase, which shows NO reductase activity within the range of the canonical membrane-bound heme b3-iron NO reductases. Using reverse-transcription polymerase chain reaction we show that although the flavorubredoxin gene (flrd) is transcribed in both aerobic and anaerobic conditions, anaerobiosis induced transcription up to 12-fold, under fermentative conditions; a 28-fold stimulation was observed in an E. coli fnr mutant strain, showing that the flavorubredoxin gene is negatively regulated by FNR. The level of anaerobic transcription was repressed three-fold by nitrate, but induced 47-fold by nitrite. The transcription factors NarL and NarP are not essential for flrd expression. Furthermore, the addition of NO within the physiological range of concentrations does not induce anaerobic transcription of flrd. Since two other E. coli proteins are known to exhibit NO reductase activity, flavohemoglobin and the pentaheme cytochrome c nitrite reductase, we have also compared the concentrations of their mRNAs with those of flavorubredoxin, under the same growth conditions. Transcription of the putative transcriptional activator of flavorubredoxin, ygaA, is also regulated by the absence of oxygen and the presence of nitrite. Levels of FlRd protein did not correlate with mRNA levels. The results reveal that a complex regulation of flavorubredoxin expression is operative, possibly by both transcriptional and post-transcriptional mechanisms.

  • Nitric oxide
  • Flavorubredoxin
  • Escherichia coli

1 Introduction

Nitric oxide (NO) plays a key role in cell physiology, both as a signalling molecule and as part of the eukaryotic defence mechanisms against pathogens [1]. Despite the important role of NO in microbial physiology, only very few prokaryotic enzymes are known to reduce NO with a significant activity. These are involved in (i) anaerobic respiration, or (ii) detoxification. The first group includes essentially the transmembranous heme b3-iron NO reductases (NOR), which are part of the prokaryotic denitrification pathway; the pentaheme nitrite reductases, involved in nitrate/nitrite respiration, mainly in the ammonification pathways have been reported also to have NO-reducing activity [2], but an accurate measurement of its activity remains to be determined. The second group is constituted only by the flavohemoglobins, which however have a low NOR activity (∼0.2 s−1). Another enzyme was recently added to this second group of NO detoxification enzymes: the Escherichia coli cytoplasmic flavorubredoxin (FlRd) [3], which was shown to have a high NOR activity, with a turnover of ∼20 s−1, a value that is within the range of the canonical membrane-bound heme b3-iron nitric oxide reductases [4]. The E. coli flavorubredoxin is a member of the family of A-type flavoproteins, which are widespread among anaerobes and facultative anaerobes. The members of this enzyme family have in common a two-domain core, built by a β-lactamase-like domain, at the N-terminal region and harboring a di-iron catalytic site, and a flavodoxin-like domain, containing one FMN moiety [5]. The enzyme from E. coli K-12, as well as from other enterobacteria, has an extra domain at the C-terminus, containing a rubredoxin-like center [3]. Downstream of the flrd gene is a gene encoding an FAD-containing NADH oxidoreductase (FlRd-red), which reduces the flavorubredoxin at the expense of NADH [3]. E. coli lacks the heme b3-iron NORs, but expresses two other enzymes capable of detoxifying NO: the flavohemoglobin (Hmp), which has a much higher activity as NO dioxygenase than as NOR [6], and the pentaheme c-type nitrite reductase (NrfA) [2]. Studies performed by measuring cell growth and NO consumption activities of whole cell extracts, using wild-type or mutant strains, provided the first evidence for the NOR activity of E. coli flavorubredoxin [7]. However, due to the presence of the above-mentioned enzymes capable of reacting with NO, and to the relative catalytic activities of each one, it becomes difficult to judge under which conditions, and at which levels, each enzyme is expressed.

Immediately upstream of the flrd/flrd-red chromosomal locus is the divergently transcribed ygaA gene, which encodes a putative σ54-dependent transcriptional regulator of E. coli [8]. YgaA strongly resembles the Ralstonia eutropha NorR protein that regulates expression of heme b3-iron nitric oxide reductase in response to NO [9]. YgaA was proposed to act as transcription activator of the flrd operon, based on the absence of significant anaerobic NO consumption activity of E. coli cells upon deletion of the ygaA gene [7], and because the product of ygaA seems to be required for the activation of the flrd promoter [10].

We now report that the genes encoding E. coli flavorubredoxin and its reductase form a di-cistronic transcriptional unit. Using reverse-transcription (RT) polymerase chain reaction (PCR) we have analyzed the levels of flrd mRNA in the wild-type E. coli K-12 grown under aerobic, anaerobic, nitrate- and nitrite-respiring conditions. The effect of exposure to NO on the transcription level was analyzed, using NO concentrations within the physiological levels and avoiding the use of other nitrosative agents that may cause additional alterations of cellular metabolism. The transcription levels of the genes encoding the other E. coli enzymes capable of detoxifying NO, the flavohemoglobin and the pentaheme nitrite reductase, as well as that of the putative transcriptional activator ygaA have been compared, and the effects of fnr, narL and/or narP mutations on the anaerobic transcription of flrd have been analyzed. Additionally, the production of FlRd under various growth conditions was also studied by immunoblotting.

2 Materials and methods

2.1 Bacterial strains and growth conditions

E. coli strains JCB3871 (fnr mutant of JCB387) [11], JCB3883 (narL mutant of JCB387) [12], JCB3875 (narP mutant of JCB387) [13], and JCB3884 (narL and narP mutant of JCB387) [13] were kindly provided by Prof. J.A. Cole. E. coli K-12 (ATCC 23716) was used as the wild-type strain. Growth was monitored by measuring culture turbidity at 600 nm (OD600) with a UV1101 Biotech Photometer WPA, and bacteria were always collected in the exponential phase. Aerobic cultures were grown at 37°C in Luria–Bertani (LB) broth [14] in phosphate buffer pH 7, in flasks filled with medium to a 1/10 ratio of total flask capacity and shaken at 200 rpm. Anaerobic cultures were cultivated at 37°C in screw-cap bottles that were filled to the top and grown either in LB medium supplemented with 20 mM glucose or in minimal salts (MS) medium pH 7 (60 mM K2HPO4, 33 mM KH2PO4, 7.6 mM (NH4)2SO4, 1.7 mM sodium citrate, 1 mM MgSO4 and 10 µM MnCl2) supplemented with 10 µg ml−1 of thiamine and 40 µg ml−1 of l-arginine, l-leucine, l-proline and l-threonine. Anaerobic cultures were grown in MS medium, using glucose as carbon source (40 mM), and 40 mM NaNO3 or 5 mM NaNO2 as electron acceptor. E. coli cells were also cultivated under growth conditions which maximize nrfA expression [2]: MS medium supplemented with fumarate (20 mM), glycerol (2 g l−1), 5% (v/v) LB, and sodium nitrite (1 mM NaNO2). Chloramphenicol (20 µg ml−1) and tetracycline (10 µg ml−1) were added where appropriate.

2.2 Bacterial exposure to NO

An NO-saturated water solution (2 mM) was prepared by flushing NO through a 1 N NaOH solution prior to its introduction into water. When cultures, grown anaerobically either in LB or in MS medium, had reached OD600 of 0.2, appropriate amounts of the NO-saturated water solution were injected to a final NO concentration of 15 µM and 150 µM. After further growth to an OD of 0.3–0.4, bacteria were harvested by centrifugation.

2.3 RNA isolation and RT-PCR experiments

The total RNA was isolated from bacteria grown as described above using RNeasy Mini and Midi Kits (Qiagen) and treated with DNase I RNase-free (Roche Applied Science). RNA concentration was evaluated spectrophotometrically at 260 nm.

Based on the E. coli K-12 genome sequence [15], forward and reverse primers were designed for PCR amplification of the following DNA fragments: 504 bp for flavorubredoxin gene (flrd1: 5′-GAAAAATAACATTCATTGGGTTGG-3′ and flrd2: 5′-CAGTAGTGTTGACCGAAAGCATC-3′); 491 bp for flavorubredoxin reductase gene (rdr1: 5′-GCAAGAGTATCGCGCCTGTGA-3′ and rdr2: 5′-CGCTAAGTTGAATCGGCTGGAG-3′); 510 bp for flavorubredoxin/flavorubredoxin reductase putative regulator gene (yga1: 5′-CCAGAACTACCGCCCGATGAATA-3′ and yga2: 5′-CGGAAAGTGTGGCGGAAAGTG-3′); 497 bp for flavohemoglobin gene (hmp1: 5′-GTTAACCGCCCATTTCTACGACC-3′ and hmp2: 5′-GATATTGCCCCGGACGGTATTC-3′), and 402 bp for cytochrome c nitrite reductase gene (nrfA1: 5′-GCTCCCGCAAAACCTGTAACTG-3′ and nrfA2: 5′-ATGGCAATCGGCACAACCTAAG-3′). Prior to RT-PCR experiments, for each pair of primers optimization of the PCR amplification conditions was established using E. coli K-12 (ATCC 23716) genomic DNA as template. RT-PCR experiments were performed using a OneStep RT-PCR Kit (Qiagen). For each RNA sample, control experiments in which the RT-PCR step was omitted confirmed the absence of any residual DNA. DNA amplification, generated in RT-PCR reactions, using different amounts of total RNA (from 10 ng to 1.0 µg), was quantified by comparison with the intensity of the 100-bp ladder DNA marker (Promega) and analyzed with Kodak Digital Science Electrophoresis Documentation and Analysis System 120. The values presented are the result of three independent RT-PCR reactions, with an average error of ∼10%.

2.4 The flrd/flrd-red operon

To prove that flrd and flrd-red genes are cotranscribed two additional primers were designed: a forward primer that hybridized in the 3′ region of the flrd gene (op1: 5′-CAGCGTCTGCCAGTGGATTTAC-3′) and a reverse primer that annealed in the 5′ region of the flavorubredoxin reductase gene (op2: 5′-AGGCGCGATACTCTTGCTGACT-3′), generating a 557-bp fragment. In addition, we have observed that the RT-PCR reactions performed with the flrd1/flrd2 and the rdr1/rdr2 primers, using the same amount of RNA, yielded DNA bands with intensities identical to those obtained with the op1/op2 primers. Hence, the latter pair was used in the subsequent RT-PCRs for testing the levels of flrd/flrd-red transcription.

2.5 Immunoblotting experiments

Polyclonal antiserum against the purified E. coli FlRd [3] was raised in rats (Eurogentec, Belgium). Cultures were grown under the same conditions used for RNA assays; cells were harvested and disrupted in a French pressure cell (SLM/Aminco), followed by centrifugation to remove cell debris. The crude extract was ultracentrifuged at 160 000×g, for 1 h, and the supernatant (soluble extract) decanted. Total soluble protein was determined according to the method of Bradford [16], using Protein Standard Mix (Sigma) for standardization. Each protein sample was separated by SDS–PAGE and transferred to nitrocellulose membranes using a semi-dry Millipore transfer system (Bio-Rad), according to the supplier's instructions. The membrane was incubated with anti-FlRd, followed by incubation with the anti-rat IgG-alkaline phosphatase conjugate (Sigma) and detected with 4-nitroblue-tetrazolium chloride and 5-bromo-4-chloro-3-indolylphosphate (Roche Applied Science) for 30 min.

3 Results and discussion

The hypothesis that the flrd and flrd-red genes form a single transcription unit [3] was proved to be correct by RT-PCR. Indeed, using a pair of appropriate primers (see Section 2), a fragment with the expected size, 557 bp, was amplified from the E. coli target mRNA.

Subsequently, RT-PCR was also used to monitor flrd and flrd-red gene amplification as a function of mRNA levels present in E. coli cultured under aerobic or anaerobic (fermentative, nitrate- and nitrite-respiring) conditions. Since E. coli contains two other gene products exhibiting NOR activity, Hmp and NrfA, their transcription levels were simultaneously evaluated.

Equal levels of gene amplification were observed for flrd, flrd-red and flrd/flrd-red (data not shown), and Fig. 1A shows that under all conditions tested the flrd/flrd-red, ygaA, hmp and nrfA genes are expressed, but the relative quantities of each mRNA are clearly dependent on the growth conditions (Table 1). For example, and independently of the growth conditions, transcription of the hmp gene could be detected using 10 ng of total RNA (data not shown). However, no transcription of the flrd/flrd-red genes was detected using the same amount of RNA, except in nitrite-grown cells. Also, in nitrate-grown E. coli, transcription of flrd/flrd-red genes was only observed using a minimum of 500 ng of RNA as template.

Figure 1

RT-PCR quantification of the transcription of E. coli flrd/flrd-red, hmp nrfA and ygaA genes. A: RNAs were isolated from E. coli K-12 grown aerobically in LB (lane 1), anaerobically in LB (lane 2), anaerobically in MS/glucose (lane 3), anaerobically in MS/40 mM nitrate (lane 4), anaerobically in MS/5 mM nitrite (lane 5), and anaerobically in MS/20 mM fumarate/1 mM nitrite (lane 6). B: Transcription levels of flrd/flrd-red, hmp, nrfA and ygaA genes in wild-type E. coli exposed to NO, grown anaerobically in LB (lane 1), anaerobically in LB with 15 µM NO (lane 2), anaerobically in LB with 150 µM NO (lane 3), anaerobically in MS/glucose (lane 4), and anaerobically in MS/glucose with 150 µM NO (lane 5). C: Transcription levels of flrd/flrd-red, hmp, nrfA and ygaA genes in E. coli wild-type and fnr-deficient strain: wild-type (lanes 1, 3 and 5) and E. coli JCB3871 (Δfnr) (lanes 2, 4 and 6) grown anaerobically in MS/glucose (lanes 1 and 2), anaerobically in MS/40 mM nitrate (lanes 3 and 4), and anaerobically in MS/5 mM nitrite (lanes 5 and 6). D: Transcription levels of flrd/flrd-red, hmp and ygaA genes in E. coli wild-type and mutant strains: wild-type (lanes 1, 5 and 9), E. coli JCB3883 (ΔnarL) (lanes 2, 6 and 10), E. coli JCB3875 (ΔnarP) (lanes 3, 7 and 11), and E. coli JCB3884 (ΔnarLP) (lanes 4, 8 and 12) grown anaerobically in MS/glucose (lanes 1–4), anaerobically in MS/40 mM nitrate (lanes 5–8), and anaerobically in MS/5 mM nitrite (lanes 9–12). The amount of total RNA used in all RT-PCR reactions was 70 ng, except for flrd/flrd-red in nitrate-grown cells (A: flrd/flrd-red, lane 4, 500 ng), ygaA grown anaerobically in MS/glucose and MS/glucose/150 µM NO (A: ygaA, lane 3 and B: ygaA, lane 5, 140 ng). M: 100-bp molecular weight marker (0.6 µg). All fragments of the marker contain 50 ng of DNA, except the 500-bp fragment (indicated with an arrow), which contains 150 ng of DNA.

View this table:
Table 1

Transcription levels of flrd/flrd-red, hmp, ygaA and nrfA genes in E. coli K-12 grown under aerobic, fermentative, nitrate- and nitrite-respiring conditions, and exposed to NO

Growth mediumDNA amplified by RT-PCR (ng)/total RNA template (ng)
Aerobic growth of E. coli K-12
Anaerobic growth of E. coli K-12
LB/NO (15 µM)3.6ndndnd
LB/NO (150 µM)13.4ndndnd
MS/glucose/NO (150 µM)7.1143.9nd3.6
  • nd, not determined.

3.1 Transcription of flrd in aerobic conditions

In aerobically grown E. coli the flrd/flrd-red mRNA abundance was low (∼60 times less) when compared with that of the hmp gene (Table 1, Fig. 1A), suggesting a minor contribution of FlRd to NO protection under aerobic conditions. Furthermore, FlRd was not detected by immunoblotting in aerobically grown E. coli (see below) which indicated a negligible amount of protein under these conditions. This was consistent with a previous report that an E. coli hmp-deficient mutant was significantly impaired by exposure to NO during aerobic growth [7]. These results are in agreement with the catalytic activity of HMP as NO dioxygenase [6], and of FlRd as NOR [4].

3.2 Transcription of flrd in anaerobic conditions

When cells were shifted from aerobic to anaerobic fermentative metabolism an approximately 12-fold increase (LB complex medium) and seven-fold increase (defined medium) in the transcription level of the flrd/flrd-red operon was measured (Table 1, Fig. 1A). These results clearly showed that the flrd/flrd-red genes were transcribed in NO-free anaerobically grown E. coli cells, and that the flrd transcription is not restricted to an NO-inducible anaerobic pathway, as previously proposed by Gardner et al. [7].

The hmp gene was also slightly more expressed in LB-fermentative compared with aerobic conditions (Table 1, Fig. 1A), in agreement with previous results [17]. In both complex and defined medium the amount of hmp mRNA was always higher (between four- and seven-fold) than that of flrd/flrd-red (Table 1, Fig. 1A). Since both genes are expressed under anaerobic conditions, it is most probable that they act together in NO protection in these conditions. In fact, Gardner et al. [7] observed that an E. coli strain deficient in either hmp or flrd, grown anaerobically and exposed to NO, did not suffer growth inhibition. Altogether, these results do not support the proposal that Hmp only detoxifies NO under aerobic conditions [18]. Our results could also suggest that Hmp would play the major role in anaerobic NO protection. However, we have shown that FlRd has a much higher NOR activity (20 s−1) [4] than Hmp (0.24 s−1) [6], i.e., the quite different activities can certainly counterbalance the different levels of transcription, thus explaining why FlRd is able to protect against NO an E. coli hmp mutant, under anaerobic conditions. In fact, only in a double hmp and flrd mutant was E. coli growth severely impaired upon NO exposure [7].

3.3 Effect of nitrate and nitrite concentration on flrd/flrd-red transcription

The effect of alternative electron acceptors, such as nitrate and nitrite, on the transcription levels of the three genes was also studied (Table 1, Fig. 1A). In agreement with a previous report [16], nitrate or nitrite induced the transcription of hmp. Also, and as expected, nitrate or nitrite respiration strongly stimulated the transcription of nrfA [19]. However, the transcription of flrd/flrd-red showed a different behavior dependent on the electron acceptor: while a strong 45-fold stimulation occurred in the presence of nitrite (Table 1, Fig. 1A), the level of anaerobic transcription of flrd/flrd-red was repressed ca. three-fold by the presence of nitrate. These results contrast with previous ones showing that, in anaerobic conditions, nitrate activated the flrd promoter while nitrite did not [10]. Although the low transcription level of flrd/flrd-red in nitrate growth conditions may suggest that FlRd contributes little to NO reduction, its participation cannot be excluded since significant FlRd expression in nitrate-grown cells of E. coli (see below) was detected.

As mentioned above, nitrite induced a strong stimulation of the transcription of the flrd operon; however, this effect is not related to a possible accumulation of NO derived from nitrite since we have also observed that NO by itself did not induce flrd transcription (see below).

3.4 Effect of NO on anaerobic transcription of flrd/flrd-red

Our results showed that transcription of hmp was strongly enhanced by the addition of NO (Table 1, Fig. 1B), as determined by Poole and coworkers [17]. In contrast, the transcription level of the flrd/flrd-red operon was unaffected by the addition of 15 µM NO to E. coli cells growing fermentatively in complex medium, and only a ca. three-fold increase was observed when the concentration was raised to 150 µM (Table 1, Fig. 1B). The possible effect of NO release caused by nitrate contamination of LB broth [20] was assessed by comparing the flrd/flrd-red transcription level in defined medium versus defined medium supplemented with 150 µM NO (Table 1). However, no significant changes were detected (Table 1) indicating that NO does not act as an effector of flrd transcription within the physiological concentrations, reported to be in the range 10–40 µM [21]. Although Hutchings et al. [10] observed an activation of the flrd promoter upon addition of sodium nitroprusside, this effect was fully dependent on the product of ygaA, which is itself insensitive to NO (see below).

3.5 FNR, NarL and NarP regulation of flrd/flrd-red transcription

FNR is known to induce the transcription of the E. coli nrfABCDEFG operon during anaerobic cell growth [22], and to negatively regulate the transcription of E. coli hmp [7]. In the fnr null E. coli strain, the nrf mRNA level decreased while the level of hmp increased (Table 2, Fig. 1C). Furthermore, the fnr mutation caused a 27-fold increase in the flrd/flrd-red transcription (Table 2, Fig. 1C), indicating that the transcription of flrd/flrd-red is negatively regulated by FNR. Analysis of the region upstream of the flrd operon shows that this regulation may be achieved by the binding of FNR to a site centered at position −177 (5′-TTGCC-N5-GTCAA-3′). However, our data showing that NO does not regulate flrd suggest that the mechanism of repression exerted by the transcription factor FNR on the flrd operon is clearly different from that on the hmp gene, which is proposed to be up-regulated by NO through FNR inhibition [17,23].

View this table:
Table 2

Comparison between the transcription levels of flrd/flrd-red, hmp and ygaA genes in E. coli wild-type and fnr, narL and/or narP null mutants, grown anaerobically

GeneStrain, genotypeDNA amplified by RT-PCR (ng)/total RNA template (ng)
flrd/flrd-redK-12, wt2.10.694
JCB3871, Δfnr57.00.7143
JCB3883, ΔnarL57.00.6200
JCB3875, ΔnarP28.50.7171.4
JCB3884, ΔnarLP57.02.4228.6
hmpK-12, wt8.634.885.7
JCB3871, Δfnr57.1142.9128.6
JCB3884, ΔnarLP143.9171.4214.3
nrfAK-12, wt2.934.873.0
JCB3871, Δfnrndnd25.7
ygaAK-12, wt4.30.471.4
JCB3871, Δfnr14.31.035.7
JCB3884, ΔnarLP32.10.671.4
  • nd, not determined.

The transcription of the flrd operon was shown to be slightly repressed by nitrate and strongly induced by nitrite. In addition, analysis of the upstream region of flrd revealed the presence of a second cis-acting element, the inverted repeat heptamer centered at −115 (5′-TACTCAT-TA-ATGGGCA-3′), that shares sequence similarity with the NarL/NarP binding site consensus. Hence, the role of the regulatory systems NarL and NarP, which integrate signals resulting from the presence of nitrate or nitrite, in flrd/flrd-red transcription was investigated.

In nitrate-respiring conditions, narL or narP mutations had no detectable effect in the flrd/flrd-red mRNA levels, while the narL narP double mutation caused a slight increase in the flrd operon transcription level (Table 2, Fig. 1D). Hence flrd is not subject to nitrate repression by narL in the presence of a high concentration of nitrate. On the other hand, the level of flrd/flrd-red mRNA almost doubled in nitrite-grown strains defective in narP and/or narL, showing that under these growth conditions NarL and NarP proteins act as negative regulators of the flrd operon. However, the increase of the transcription level of flrd in the strains defective in narL and/or narP is much smaller when compared with the induction caused by the addition of nitrite. Furthermore, the levels of flrd/flrd-red mRNA were significantly increased in narL and/or narP mutant strains grown anaerobically in the absence of nitrate or nitrite. Thus, the results suggest that NarL and NarP proteins are competing with the main regulator factor of flrd, which is activated by nitrite (or by a derivative product), and may bind to the nucleotide region in the vicinity of the putative narL/narP binding site. The binding of FNR and NarL/NarP regulators most probably occurs in the nucleotide region that encodes ygaA, which is transcribed in the opposite direction. This proposal is corroborated by the failure of ygaA to complement the ygaA mutation in trans, which already suggested the existence of a cis-acting sequence within the ygaA coding region required for flrd transcription [10].

3.6 Effect of growth conditions on the transcription level of ygaA

In E. coli, ygaA is adjacent to flrd; the two genes are divergently transcribed and the ygaA and flrd operon promoters overlap. Recent data led to the proposal that the ygaA gene product is involved in the regulation of flrd [7,10]. Hence, for further characterization of this gene we determined ygaA mRNA levels under the same growth conditions tested for flrd/flrd-red.

In aerobically grown E. coli low levels of ygaA mRNA were measured (Table 1, Fig. 1A), but when the metabolism was shifted to anaerobic conditions, the transcription of ygaA increased about 30-fold. The repressor effect of nitrate on ygaA transcription was stronger (∼10-fold) than that observed for flrd/flrd-red, and nitrite elevated the anaerobic ygaA mRNA level 17-fold (Table 1). However, the hypothesis that NO formed by nitrite reduction could be the actual signal molecule was ruled out since the ygaA mRNA level showed a negligible decrease upon addition of NO (150 µM) (Table 1, Fig. 1B). Hutchings et al. [10] reported that transcription at the ygaA promoter was slightly inhibited by sodium nitroprusside, a result that fully agrees with our data. Although these authors observed that the flrd promoter was activated by sodium nitroprusside (100 µM), they also detected that flrd activation was abolished in a flrd/flrd-red mutant. Therefore, it was concluded that the true signalling compound recognized by ygaA is not NO, but a compound derived from the interaction of FlRd/FlRd-red proteins with reactive nitrogen species [10].

The effect on ygaA transcription of the FNR regulator and of the nitrate and nitrite mediating response regulators NarL and NarP was also examined (Fig. 1C and 1D). We observed that in anaerobic conditions the transcription of ygaA is slightly repressed by FNR, but the utilization of nitrite as the final electron acceptor reverses this behavior (Table 2). The narL and narP null alleles had no effect on ygaA transcription in cells grown with nitrate or nitrite, but increased the ygaA mRNA level in cells grown fermentatively (Table 2). These results showed that ygaA transcription responds to two signals, the absence of oxygen and the presence of nitrite, in a way that very much resembles the regulation of the flrd operon, suggesting that the same control elements are mediating the induction of flrd and ygaA.

3.7 Effect of growth conditions on FlRd expression

To complement the transcription data, the FlRd expression level in wild-type and mutant strains of E. coli grown under different conditions was evaluated by immunoblotting. The results (Fig. 2) showed that there is no direct correlation between the transcription level and the protein level, although the same trend was observed in some cases. In E. coli grown aerobically in rich medium no protein expression could be detected, even when 200 µg of soluble protein extract was analyzed (Fig. 2A, lane 2 and Fig. 2B, lane 6). Only during anaerobic growth was expression of flavorubredoxin observed (Fig. 2A, lanes 3 and 4). Addition of 150 µM NO to the defined medium of anaerobically grown wild-type E. coli increased flrd transcription, with a higher effect in complex medium (Table 1), and the same trend was observed for the FlRd protein level (Fig. 2A, lanes 7 and 8). As observed for the mRNA levels (Table 2), narLP-defective E. coli expressed a higher amount of FlRd in nitrite-respiring conditions than in nitrate-respiring conditions (Fig. 2B, lanes 4 and 5). The difference between the flrd mRNA levels in the E. coli fnr minus strain grown fermentatively and in nitrite-respiring conditions (Table 2) was much smaller than the difference between the protein expressed under the same conditions (Fig. 2B, lanes 2 and 3). Also, the relative levels of FlRd protein in fermentative, nitrate-, and nitrite-grown cells of wild-type E. coli were inconsistent with the transcription levels. While a much higher flrd transcription level was observed when E. coli wild-type was grown in nitrite-respiring conditions compared to other growth conditions, FlRd expression in nitrite-grown E. coli was similar to that in fermentatively grown E. coli and lower than in nitrate-grown E. coli (Fig. 2A, lanes 4–6). These results were growth phase-independent since in cells collected at the stationary phase the relative amounts of FlRd detected were similar (Fig. 2B, lanes 7–9). The discrepancy between the mRNA level and protein abundance could be due to a different turnover rate of flrd mRNA caused by the different growth conditions or to post-transcriptional control. However, it was recently shown that, at least between nutrient-rich and defined media, no major variations in the mRNA half-lives of the E. coli genes were observed [24]. Therefore, the possibility of a post-transcriptional control that together with the transcription regulation acts on FlRd expression should be considered.

Figure 2

Immunodetection of FlRd protein with an antibody raised against E. coli flavorubredoxin. A: Soluble extracts of wild-type E. coli K-12 grown aerobically in LB (lane 2), anaerobically in LB (lane 3), anaerobically in minimal salts (MS) (lane 4), anaerobically in MS with nitrate (lane 5), anaerobically in MS with nitrite (lane 6), anaerobically in LB/150 µM NO (lane 7) and anaerobically in MS/150 µM (lane 8). B: Soluble extracts of E. coli JCB3871 (Δfnr) grown anaerobically in MS (lane 2), E. coli JCB3871 (Δfnr) grown anaerobically in MS with nitrite (lane 3), E. coli JCB3884 (ΔnarLP) grown anaerobically in MS with nitrate (lane 4) and E. coli JCB3884 (ΔnarLP) grown anaerobically in MS with nitrite (lane 5). E. coli cells used in A and B (lanes 2–5) were grown and collected under the same conditions used for the RNA assays and 50 µg of protein was used for the SDS gel electrophoresis. Western blots of wild-type E. coli collected in the stationary phase were analyzed using always 100 µg of total protein extracted from cells grown fermentatively (B, lane 7), anaerobically with nitrate (B, lane 8) and anaerobically with nitrite (B, lane 9), except from cells grown aerobically in LB (B, lane 6) for which 200 µg of total protein was used. Lanes 1 (A, B), 0.2 µg of purified E. coli flavorubredoxin.

3.8 Final conclusions

The results obtained by studying directly the flrd transcription in wild-type and mutant strains of E. coli are summarized in Fig. 3. We have shown that the newly discovered E. coli NOR, FlRd, is essentially expressed during anoxic growth and its transcription is only affected by non-physiological NO concentrations. In fermentative conditions, although its transcription level is lower than that of flavohemoglobin, the much higher NO-reducing activity of FlRd can fully compensate its lower abundance under these conditions. We have also observed that anaerobic transcription of the flrd operon is markedly induced by nitrite, but not by nitrate. In addition, we could conclude that the potential FNR and NarL/NarP boxes, located in the upstream region of flrd, may indeed bind the respective transcription factors since variations in the flrd mRNA level were observed when null alleles of fnr, narL and/or narP were transduced into E. coli. Nevertheless, our results strongly suggest that another factor, whose nature remains to be elucidated, is playing the major control, although in competition with FNR and NarL/NarP, and its action is mainly triggered by the presence of nitrite. Furthermore, we observed that the transcription of the ygaA gene, proposed to act as a transcriptional activator of flrd, is also regulated by the absence of oxygen and the presence of nitrite in the culture medium of E. coli. In fact, expression of FlRd cannot be simply related to YgaA, since we have shown that a complex transcription regulation and, most probably, a post-transcriptional mechanism control its expression.

Figure 3

Comparative measurement of flrd transcription. Strains: wild-type E. coli is indicated as K-12, E. coli JBC3871 fnr null mutant as Δfnr, E. coli JBC3883 narL null mutant as ΔnarL, E. coli JBC3875 narP null mutant as ΔnarP, and E. coli JBC3884 double narLP null mutant as ΔnarLP. Growth conditions: all strains were cultivated anaerobically in rich medium (LB), minimal salts medium (MS), supplemented with nitrate or nitrite or grown in fumarate/nitrite, as described in Section 2. NO(15) and NO(150) represent addition of 15 µM or 150 µM of nitric oxide, respectively.

Finally, note that the homologues of E. coli FlRd present in the sulfate-reducing bacterium Desulfovibrio gigas [25], in Rhodobacter capsulatus [26] and in Methanobacterium thermoautotrophicum [27] are also expressed under anaerobic conditions, in the absence of NO. Although the NO-reducing activity of these enzymes is still unknown, the high degree of amino acid similarity of these enzymes, especially the conservation of the ligands to the di-iron catalytic site, strongly suggests that they may play a similar role in NO detoxification. This role would be particularly important in prokaryotes lacking other NO-reducing enzymes, such as the methanogens.


We thank Prof. J.A. Cole for supplying the E. coli bacterial strains defective in the fnr, narL, narP and narLP genes and for helpful discussions. P.N.d.C. is the recipient of a grant from Praxis XXI program (BPD/6958/2001). This work was supported by the Fundação da Ciência e Tecnologia Project POCTI/1999/BME/36558 (M.T.).


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
  26. [26].
  27. [27].
View Abstract