Regulation by Nucleoid-Associated Proteins at the Escherichia coli nir Operon Promoter

Abstract

The Escherichia coli K-12 nir operon promoter can be fully activated by binding of the regulator of fumarate and nitrate reduction (FNR) to a site centered at position −41.5 upstream of the transcript start, and this activation is modulated by upstream binding of the integration host factor (IHF) and Fis (factor for inversion stimulation) proteins. Thus, transcription initiation is repressed by the binding of IHF and Fis to sites centered at position −88 (IHF I) and position −142 (Fis I) and activated by IHF binding to a site at position −115 (IHF II). Here, we have exploited mutational analysis and biochemistry to investigate the actions of IHF and Fis at these sites. We show that the effects of IHF and Fis are position dependent and that IHF II functions independently of IHF I and Fis I. Using in vitro assays, we report that IHF and Fis repress transcription initiation by interfering with RNA polymerase binding. Differences in the upstream IHF and Fis binding sites at the nir promoter in related enteric bacteria fix the level of nir operon expression under anaerobic growth conditions.

The Escherichia coli nir operon encodes a cytoplasmic NADH-dependent nitrite reductase, which is expressed under anaerobic conditions and is responsible for the reduction of nitrite to ammonium ions (19). Transcription of this operon is driven from a single promoter (pnir) and is induced by the fumarate and nitrate reduction (FNR) protein, a global transcription regulator that activates the expression of many genes in response to oxygen starvation (9, 11). At pnir, FNR binds to a DNA site centered at position −41.5 and activates transcription directly by interacting with RNA polymerase (25) (Fig. 1A). This activation is modulated by the upstream binding of nucleoid-associated proteins, integration host factor (IHF) and Fis (factor for inversion stimulation) (3), two DNA binding proteins that play a role in shaping the folded bacterial chromosome (1). IHF binds to two sites at pnir: at IHF I, centered at position −88, and at the lower-affinity IHF II, centered at position −115 (Fig. 1A). Binding of IHF to IHF I and to IHF II has opposing effects, repressing and activating pnir expression, respectively (3, 5). Binding of Fis to an upstream site centered at position −142, Fis I, represses FNR-dependent transcription. Additionally, Fis binds independently to a weaker downstream site (Fis II, centered at +23), which also represses nir expression (3, 5, 28).

FIG. 1.

Organization of the nir promoter region. (A) The figure shows a schematic representation of the nir promoter fragments used in this work. The upstream boundary for each promoter fragment is given, and the location of the pnir transcription start site is indicated by a bent arrow. FNR and NarL/NarP binding sites are represented by inverted arrows, while IHF and Fis binding sites are depicted by boxes. The transcription start site is denoted +1, and the location of the p99G, p112G, and p146A substitutions, which disrupt the IHF I, IHF II, and Fis I binding sites, are indicated (3, 5, 28). The sites where DNA was inserted into pnir promoter fragments at positions −60, −106, and −134 are labeled 1, 2, and 3, respectively. (B) The figure illustrates measured β-galactosidase activities of JCB3884 cells carrying pRW50 containing the different nir promoter fragments described in panel A. Cells were grown aerobically and anaerobically in minimal salts medium plus 0.4% glucose. β-Galactosidase activities are expressed as nmol of ONPG hydrolyzed min⁻¹ mg⁻¹ of dry cell mass. Each activity is the average of three independent determinations.

Anaerobic expression of the nir operon can be further induced by the nitrite or nitrate ions. This is achieved by two homologous transcription factors, NarL and NarP, which are phosphorylated by the sensor kinases NarX and NarQ (23). Once phosphorylated, NarL or NarP binds to the same target site centered at position −69.5 (Fig. 1A) and activates transcription by displacing IHF from IHF I, thereby counteracting the repression by IHF and Fis from upstream sites (3, 25, 28).

The mechanisms by which IHF and Fis modulate the activity of different promoters are still poorly understood. Thus, in this work we have exploited mutational analysis in vivo and biochemical studies in vitro to investigate the roles of IHF and Fis at the E. coli K-12 nir operon promoter. We also report that variation of the upstream nir promoter sequences in related enteric bacteria set different levels of activity in response to the formation of different IHF-promoter DNA complexes.

MATERIALS AND METHODS

Bacterial strains, plasmids, and DNA fragments.

Bacterial strains, plasmids, and promoter fragments used in this work are listed in Table 1, and oligonucleotides are listed in Table 2. Standard methods for cloning and manipulating DNA fragments were used. By convention, locations at the nir promoter are labeled with the transcript start point as +1, and upstream and downstream locations are prefixed with minus and plus signs, respectively. Our starting point was the pnir7150 fragment that carries the pnir sequence from position −150 to position +36, with an upstream EcoRI linker and a downstream HindIII linker and with an NsiI site at position −60 (Fig. 1A). Single base substitutions in pnir are denoted pNX, where N is the position of the substitution relative to the transcript start, and X is the substituted base in the nontemplate strand. For routine DNA manipulations and as a source of fragments for gel retardation and footprinting analysis, fragments were cloned into plasmids pAA121 or pSR. To measure promoter activities, fragments were cloned into the lac expression vector pRW50. Derivatives of pAA121 and pSR were maintained in host cells using medium supplemented with 100 μg ml⁻¹ ampicillin, while pRW50 derivatives were maintained with 15 μg ml⁻¹ tetracycline.

TABLE 1.

Bacterial strains, plasmids, and promoter fragments used in this work

Strain, plasmid, or promoter	Relevant characteristic(s)	Reference or source
Strains
E. coli K-12 strains
JCB387	Δnir Δlac	18
JCB3884	JCB387 narL narP253::Tn10dCm	25
JCB38841	JCB3884 fis985 str/spcr	28
JCB38849	JCB3884 himA452::Tn10dTc	3
JCB38849S	Tets version of JCB38849 isolated as a fusaric acid-resistant colony	This work
JRG1728	Δfnr Δlac	26
E. coli E2348/69	EPEC	I. Henderson
S. enterica serovar Typhimurium LT2	Wild type	I. Henderson
Plasmids
pAA121	Cloning vector for EcoRI-HindIII fragments derived from pBR322	12
pSR	pBR322 derivative containing a λ oop transcription terminator	13
pRW50	Broad-host-range lacZ fusion vector for cloning promoters on EcoRI-HindIII fragments; contains the RK2 origin of replication	14
Promoters
pnir7150	E. coli nir promoter fragment carrying nucleotide sequences from −150 to +36	2
pnir7133	E. coli nir promoter fragment carrying nucleotide sequences from −133 to +36	This work
pnir7106	E. coli nir promoter fragment carrying nucleotide sequences from −106 to +36	This work
pnir7083	E. coli nir promoter fragment carrying nucleotide sequences from −83 to +36	This work
pnir7150/p146A	E. coli pnir7150 promoter fragment carrying a T-to-A substitution at position −146	28
pnir7150/p112G	E. coli pnir7150 promoter fragment carrying a C-to-G substitution at position −112	5
pnir7150/p99G	E. coli pnir7150 promoter fragment carrying an A-to-G substitution at position −99	28
pnir7150/p99Gp146A	E. coli pnir7150 promoter fragment carrying the p99G and p146A substitutions	28
pnir7150/p99Gp112Gp146A	E. coli pnir7150 promoter fragment carrying the p99G, p112G, and p146A substitutions	5
pnir7150/H74.5	E. coli pnir7150 promoter fragment carrying 5-bp insertion at position −60	28
pnir7150/+5	E. coli pnir7150 promoter fragment carrying 5-bp insertion at position −134	This work
pnir7150/+10	E. coli pnir7150 promoter fragment carrying 10-bp insertion at position −134	This work
pnir7133/+5	E. coli pnir7133 promoter fragment carrying 5-bp insertion at position −106	This work
pnir7133/+10	E. coli pnir7133 promoter fragment carrying 10-bp insertion at position −106	This work
pnir7106/+5	E. coli pnir7106 promoter fragment carrying 5-bp insertion at position −60	This work
pnirEPEC	EPEC nir promoter fragment carrying nucleotide sequences from −150 to +36	This work
pnirSTM	S. enterica serovar Typhimurium nir promoter fragment carrying nucleotide sequences from −150 to +36	This work
pnrf53/Δ61	E. coli nrf promoter fragment carrying nucleotide sequences from −61 to +131	4
pnir-nrf	Promoter fragment carrying pnir7150 nucleotide sequences from +150 to −61 and pnrf53 nucleotide sequences from −60 to +131	25
FF(−41.5)	Semisynthetic FNR-dependent promoter carrying a consensus FNR site at position −41.5	26

TABLE 2.

Oligonucleotide primers

Name	Sequencea
D4600	5′-GTAGTCGGTGTGTTCAC-3′
D5431	5′-ACCTGACGTCTAAGAAACC-3′
D10520	5′-CCCTGCGGTGCCCCTCAAG-3′
D10527	5′-GCAGGTCGTTGAACTGAGCCTGAAATTCAGG-3′
nir7083	5′-CCCGAATTCCCGGGGATCCCGCAATATACCCATTAAGGAG-3′
nir7106a	5′-CCCGAATTCCCGGATCCTTAAGARAATTTATACAAATC-3′
nir7133	5′-CCCGAATTCAACATGAAATATCAGACAATT-3′
nir7133+5a	5′-CAGACAATTSCGTGACTTAAGTTAAGAAAATTTATACAAATCAGC-3′
nir7133+10a	5′-CAGACAATTSCGTGACTTAAGTTCAGTTAAGAAAATTTATACAAATCAGC-3′
nir7150+5a	5′-CCCGAATTCCCTGTCWATTTTTTGCACAAACTGAACATGAAATATCAGACAATT-3′
nir7150+10a	5′-CCCGAATTCCCTGTCWATTTTTTGCACAAACTCAATTGAACATGAAATATCAGACAATT-3′
nirp112G	5′-CTTAAGTCACGCAATTGTCTG-3′
nirUPEPEC	5′-CCCGAATTCCCTGTCTGTTTTTTGCACAAAC-3′
nirUPSTM	5′-CCCGAATTCCCTGTATGTTATTTGTACAAAC-3′
nirDOWN	5′-CCCAAGCTTGGACTTTGCTCATTTTTGCC-3′

^aR is A or G, S is C or G, and W is A or T.

Construction of pnir deletion and hybrid promoter fragments.

Derivatives of the starting pnir7150 fragment were constructed and cloned as EcoRI-HindIII fragments into pRW50. The pnir7083 promoter fragment (pnir sequences from −83 to +36) was generated by PCR using primers nir7083 and D4600 and pAA121/pnir7150 as a template. The pnir7106 promoter fragments (pnir sequences from −106 to +36) were constructed using PCR. Primers nir7106 and D10527 were used to amplify the nir promoter region from pRW50 plasmids containing various pnir7150 promoter derivatives. The pnir7106/+5 fragment, in which a 5-bp insertion was introduced at position −60, was generated by using pRW50/pnir7150/H74.5 as a template. The pnir7133 promoter fragment (pnir sequences from −133 to +36) was generated by PCR using primers nir7133 and D4600 and pAA121/pnir7150 as a template. The p99G and p112G substitutions were introduced into pnir7133 by using versions of pnir7150 that carried the relevant substitution(s) as a template. The pnir7133/+5 and pnir7133/+10 fragments, which introduce 5- and 10-bp insertions at position −106, were generated using megaprimer PCR. Promoter DNA was amplified using primer D10527 and either primer nir7133+5 or nir7133+10 with pRW50/pnir7106 as the template. The purified PCR products were then used in a second round of PCR with the primer D10520 and pRW50/pnir7133. The pnir7150/+5 and pnir7150/+10 promoter fragments, which carry 5- and 10-bp insertions, respectively, at position −134, were constructed using PCR. The nir promoter was amplified from pRW50/pnir7133 using primer D10527 and either primer nir7150+5 or primer nir7150+10. Megaprimer PCR was used to combine the p112G substitution, carried by pnir7150, with the p146A or p99G substitutions. The upstream region of pnir was initially amplified using primer D5431 and primer nirp112G with either pAA121/pnir7150 or pAA121/pnir7150/p146A as the template. Purified PCR products were then used in a second PCR round with the primer D4600, using pAA121/pnir7150 or pAA121/pnir7150/p99G as the template. The pnir-nrf fusion promoters were constructed by cloning pnir7150 EcoRI-NsiI fragments, carrying different substitutions, into pAA121/pnir-nrf restricted with EcoRI-NsiI.

Construction of nir promoter fragments from other enteric bacteria.

The DNA sequences of the nir operon promoter from other enteric bacteria were compiled from the xBASE (http://xbase.bham.ac.uk/colibase) online database (8). DNA fragments carrying sequence from position −150 to +36 at the nir promoter of enteropathogenic E. coli (EPEC) and Salmonella enterica serovar Typhimurium were amplified by PCR using the primers nirUPEPEC and nirUPSTM, respectively, with the primer nirDOWN. PCR products were restricted with EcoRI and HindIII and cloned into pRW50.

Assays of nir promoter activity.

To assay expression from pnir derivatives cloned into the lac expression vector pRW50, different host strains were transformed, and β-galactosidase activity was measured as described by Jayaraman et al. (10) using the Miller protocol (15). Cells were grown in minimal medium (minimal salts with 0.4% glycerol, 10% Lennox broth, 40 mM fumarate) supplemented with 0.4% glucose at 37°C (20). For aerobic growth, cells were shaken vigorously, while for anaerobic growth they were held static in growth tubes (150 mm long and 15 mm in diameter). Aerobic cultures were grown to an optical density at 650 nm of 0.2 to 0.3; anaerobic cultures were grown to an optical density at 650 nm of 0.4 to 0.6 and assayed as described previously (3). β-Galactosidase activities are reported as nmol of ONPG (o-nitrophenyl-β-d-galactopyranoside) hydrolyzed under our assay conditions min⁻¹ mg⁻¹ of dry cell mass, and each activity is the average of three independent determinations.

Proteins.

Overexpression and purification of FNR protein containing the DA154 substitution that renders FNR active under aerobic conditions were performed as described in Wing et al. (27). Purified IHF protein was prepared by the method of Nash et al. (16), and purified Fis protein was donated by Rick Gourse and prepared according to Osuna et al. (17). RNA polymerase holoenzyme was purchased from Epicentre Technologies.

Gel retardation assays.

Gel retardation assays were carried out as described in Browning et al. (4). Purified EcoRI-HindIII fragments carrying pnir were end labeled with [γ-³²P]ATP, and ∼0.5 ng of fragment was incubated with various amounts of different proteins. The reaction buffer contained l0 mM potassium phosphate (pH 7.5), 100 mM potassium glutamate, 1 mM EDTA, 50 μM dithiothreitol, 5% glycerol, 500 μg ml⁻¹ bovine serum albumin, and 25 μg ml⁻¹ herring sperm DNA. The final reaction volume was 10 μl. After incubation at 37°C for 20 min, samples were electrophoresed in 0.25× Tris-borate-EDTA buffer on a 6% polyacrylamide gel (12 V cm⁻¹) containing 2% glycerol. Gels were analyzed using a Bio-Rad Molecular Imager FX and Quantity One software (Bio-Rad). In experiments to examine the binding of RNA polymerase to DNA fragments carrying pnir, samples were incubated at 37°C for 30 min and separated using 5% polyacrylamide gels.

DNase I and potassium permanganate footprinting experiments.

DNase I and potassium permanganate footprinting experiments were performed on ³²P-end-labeled pnir fragments, using the protocols of Savery et al. (22). Each reaction mixture (20 μl) contained a final concentration of 1.35 nM template DNA. The buffer composition was 20 mM HEPES (pH 8.0), 5 mM MgCl₂, 50 mM potassium glutamate, 1 mM dithiothreitol, 500 μg ml⁻¹ bovine serum albumin, and 25 μg ml⁻¹ herring sperm DNA. For potassium permanganate footprinting reactions, herring sperm DNA was omitted, and E. coli RNA polymerase holoenzyme (Epicentre Technologies) was included at a final concentration of 50 nM. Samples were analyzed by electrophoresis on denaturing gels, calibrated with Maxam-Gilbert G+A sequencing reactions. For quantification, a Bio-Rad Molecular Imager FX was used with Bio-Rad Quantity One software.

RESULTS

Modulation of the nir promoter from upstream sites.

Previously, we investigated Fis- and IHF-mediated regulation at pnir using the pnir7150 fragment (nir sequences from −150 to +36) (Fig. 1A; Table 1). By disrupting Fis I, IHF II, and IHF I using the p146A, p112G, and p99G substitutions, respectively (Fig. 1A), we showed that FNR-dependent transcription at pnir is repressed by Fis I and IHF I and stimulated by IHF II (3, 5). To investigate how IHF and Fis modulate FNR-dependent activation of pnir, we generated a set of nested deletions (pnir7133, pnir7106, and pnir7083) that sequentially remove the upstream Fis I, IHF II, and IHF I sites (Fig. 1A). The resulting fragments were cloned into the lac expression vector pRW50 to generate pnir-lac fusions, and measurements of promoter activity were made in the Δlac narL narP strain, JCB3884, growing anaerobically in medium without added nitrite or nitrate ions. Thus, our measurements reflect the ability of FNR to activate pnir without the aid of NarL or NarP. Data illustrated in Fig. 1B show that, as expected, pnir activity was increased by deletion of Fis I. Activity was then decreased by deletion of IHF II but restored by further deletion of IHF I.

Effects of IHF are different at the nir promoter in EPEC and S. enterica serovar Typhimurium.

We examined the organization of pnir in pathogenic E. coli strains and related enteric bacteria by comparing different pnir sequences. The alignment in Fig. 2 shows that while core promoter sequences (positions −60 to +36) are identical, differences occur in the upstream region (positions −150 to −60), especially within the Fis and IHF binding sites. To measure the effects of these differences, we cloned the nir promoter from EPEC (fragment pnirEPEC) and S. enterica serovar Typhimurium (fragment pnirSTM) into pRW50 and measured activity in strain JCB3884 (narL narP) under anaerobic conditions. The results illustrated in Fig. 3A show that the activity of the EPEC nir promoter is lower than that of the E. coli K-12 promoter, while the activity of the Salmonella promoter is higher. Thus, the differences in promoter sequence do influence anaerobic gene expression.

FIG. 2.

Alignment of the nir promoter sequences from different enteric bacteria. The figure shows the sequence of the E. coli K-12 (EC K-12) (NC00913) pnir promoter from position −150 to +36 aligned with the corresponding nir promoter regions from enterohemorrhagic E. coli (EHEC) (NC002695), uropathogenic E. coli (UPEC) (NC004431), enteroaggregative E. coli (EAEC), EPEC, Shigella flexneri (SFX) (NC004741) and S. enterica serovar Typhimurium (STM) (NC003197). The location of the transcription start site for pnir is indicated by lowercase text, and the −10 element is bold and underlined. The locations of FNR and NarL/NarP binding sites are represented by inverted arrows, while IHF and Fis binding sites are depicted by boxes. Differences between the E. coli K-12 sequence and other promoters are highlighted by black boxes.

FIG. 3.

Expression of nir operon promoters from different enteric bacteria. The figure shows the β-galactosidase activities of JCB3884 (narL narP) (A), JCB38841 (narL narP fis) (B), and JCB38849S (narL narP ihfA) (C) cells carrying pRW50 containing pnir fragments from E. coli K-12 (EC K-12), EPEC, and S. enterica serovar Typhimurium (STM), as indicated. Cells were grown aerobically and anaerobically in minimal salts medium plus 0.4% glucose as indicated. In panels B and C, only anaerobic measurements are shown. β-Galactosidase activities are expressed as nmol of ONPG hydrolyzed min⁻¹ mg⁻¹ of dry cell mass, and each activity is the average of three independent determinations. The relative increase in repression due to Fis or IHF for each promoter is shown in parentheses.

To investigate whether the differences are due to Fis or IHF, the expression from each promoter was compared in fis and ihfA derivatives of strain JCB3884. Note that IHF and Fis proteins from EPEC and S. enterica serovar Typhimurium are either identical to those from E. coli K-12 or differ by only one amino acid residue. Results in shown in Fig. 3B indicate that anaerobic expression from all promoters was increased in the fis strain by approximately fourfold, indicating that Fis represses each promoter similarly. However, in the ihfA strain (Fig. 3C), anaerobic expression from pnirEPEC and pnirSTM was increased by different amounts (6- and 2.9-fold, respectively) in comparison to the 3.9-fold for the E. coli K-12 promoter. This indicates that IHF is responsible for setting alternative levels of anaerobic expression, with IHF repressing the EPEC promoter more and the Salmonella promoter less. Note that these effects are specific to the nir promoter constructs as the promoter activity of the semisynthetic FNR-dependent promoter, FF(−41.5) (which carries a consensus FNR site at position −41.5), was unaffected by the genetic background. Additionally, all promoters were dependent on FNR since anaerobic expression was negligible in the JRG1728 fnr strain (data not shown).

Gel retardation assays were performed using purified IHF and labeled nir promoter fragments. Incubation of purified IHF with the E. coli K-12 promoter fragment resulted in two retarded species (Fig. 4). The first IHF-DNA complex, which formed at lower IHF concentrations, corresponds to IHF binding at IHF I while the second, more mobile, species appeared when both IHF sites were occupied (5). Two distinct IHF-DNA complexes were also observed with the pnirEPEC fragment (Fig. 4A); however, the first IHF-DNA complex appeared at a lower IHF concentration and was more retarded than the second. This indicates that IHF binds to IHF I at the EPEC promoter with higher affinity and bends the DNA more. With the pnirSTM fragment, three complexes were detected (Fig. 4B). The first retarded complex, in which IHF I was occupied (see below), had increased mobility compared to the corresponding complex with the E. coli K-12 pnir fragment, suggesting that IHF bends pnirSTM less when bound at IHF I. At higher concentrations, two further complexes were observed. The most abundant species results when both IHF sites are occupied, while the more retarded complex is most likely due to binding of IHF at IHF II alone. DNase I footprint analysis confirmed that, at each promoter, IHF I was occupied first, with IHF II being bound at higher concentrations (Fig. 5). Thus, we conclude that IHF forms different IHF-DNA complexes when it binds to the EPEC and Salmonella promoters.

FIG. 4.

Gel retardation assays of the nir promoter. The figure shows gel retardation assays of pnir fragments from E. coli K-12 (EC K-12), EPEC, and S. enterica serovar Typhimurium incubated with purified IHF. (A) End-labeled pnir fragments were incubated with increasing concentrations of purified IHF protein: lanes 1 to 7, pnirEC K-12 EcoRI-HindIII fragment; lanes 8 to 14, pnirEPEC EcoRI-HindIII fragment. The concentration of IHF protein in each reaction mixture was as follows: lanes 1 and 8, no protein; lanes 2 and 9, 6 nM; lanes 3 and 10, 13 nM; lanes 4 and 11, 25 nM; lanes 5 and 12, 50 nM; lanes 6 and 13, 100 nM; lanes 7 and 14, 300 nM. (B) End-labeled pnir fragments were incubated with increasing concentrations of purified IHF protein: lanes 1 to 5, pnirEC K-12 EcoRI-HindIII fragment; lanes 6 to 10, pnirSTM EcoRI-HindIII fragment. The concentration of IHF protein in each reaction mixture was as follows: lanes 1 and 6, no protein; lanes 2 and 7, 25 nM; lanes 3 and 8, 50 nM; lanes 4 and 9, 100 nM; lanes 5 and 10, 300 nM. On both panels, dotted lines highlight the differences in mobility of the retarded complex due to the binding of IHF to IHF I for each promoter fragment.

FIG. 5.

DNase I footprint analysis of nir promoter fragments. (A) The figure shows in vitro DNase I footprint experiments with purified IHF. End-labeled pnir fragments were incubated with increasing concentrations of IHF protein and subjected to DNase I footprinting: lanes 1 to 7, pnir fragment from E. coli K-12 (pnirEC K-12); lanes 8 to 14, pnirEPEC; lanes 15 to 21, pnirSTM. The concentration of IHF in each reaction mixture was as follows: lanes 1, 8, and 15, no protein; lanes 2, 9, and 16, 15 nM; lanes 3, 10, and 17, 30 nM; lanes 4, 11, and 18, 59 nM; lanes 5, 12, and 19, 117 nM; lanes 6, 13, and 20, 234 nM; and lanes 7, 14, and 21, 469 nM. Gels were calibrated using Maxam-Gilbert G+A (lane GA) sequencing reactions of the labeled fragment. (B) Quantification of IHF binding to the pnirSTM promoter fragment. The binding of IHF to pnirSTM was analyzed using data from lanes 15 and 18 in panel A and Quantity One software (Bio-Rad). Boxes indicate the location of IHF sites, and selected positions are shown. The DNase I cleavage site at position −120 within IHF II, which is unaltered by the addition of IHF, is starred.

pnir upstream sequences are a functional unit that can be transplanted into another promoter.

Expression from the E. coli K-12 nrf promoter (pnrf) is induced by the binding of FNR to a site centered at position −41.5 (4, 6, 7, 25). To investigate whether upstream sequences from pnir could regulate another FNR-dependent promoter, the nir upstream sequence from position −150 to −61 was fused to the nrf promoter sequence from positions −60 to +131 to generate the pnir-nrf fusion promoter. Derivatives carrying point mutations in Fis I (p146A), IHF II (p112G), and IHF I (p99G) were also constructed, and each promoter fragment was cloned into pRW50. Results illustrated in Fig. 6 show that expression from the pnrf53/Δ61 promoter (containing nrf sequences from position −61 to +131) without upstream nir sequences was induced during anaerobic growth, but this induction was greatly suppressed by the introduction of upstream nir sequences in the pnir-nrf fusion promoter. Point mutations in Fis I (p146A) or IHF I (p99G) partially relieved this suppression, while disruption of IHF II (p112G) enhanced it. In each case, anaerobic induction was dependent on FNR since expression was negligible in the JRG1728 fnr strain (data not shown).

FIG. 6.

Upstream sequences from pnir can regulate the nrf promoter. The figure illustrates measured β-galactosidase activities of JCB3884 (narL narP) cells carrying pnrf/Δ61 and pnir-nrf promoter fragments subcloned into pRW50. The pnrf53/Δ61 fragment contains nrf sequences from position −61 to +131. Substitutions were introduced into the pnir-nrf fragment to alter the DNA sites for Fis I (p146A), IHF I (p99G), and IHF II (p112G) from pnir. Cells were grown aerobically and anaerobically in minimal salts medium. β-Galactosidase activities are expressed as nmol of ONPG hydrolyzed min⁻¹ mg⁻¹ of dry cell mass. Each activity is the average of three independent determinations that varied by less than 10%.

Upstream IHF and Fis sites must be correctly positioned to regulate pnir.

We have exploited our upstream nested deletions to investigate how IHF and Fis binding at each site modulates FNR-dependent activation at pnir. We used the pnir7106 fragment and the pnir7106/+5 derivative with a 5-bp insertion at position −60 to investigate the effects of IHF at IHF I (Fig. 1A). Results in Table 3 show that the 5-bp insertion increased anaerobic expression. The p99G substitution that disrupts IHF binding to IHF I increased expression from pnir7106 but not from the pnir7106/+5 promoter. Thus, IHF I must be correctly positioned for repression.

TABLE 3.

β-Gal activities of JCB3884 cells carrying different pnir7106 fusions

Promoter fragmenta	Mutated site	β-Gal activity (nmol of ONPG hydrolyzed min−1 mg−1 of dry cell mass)b		% Differencec
		+ O2	−O2
pnir7106		20	700
pnir7106/p99G	IHF I	30	2,600	370
pnir7106/+5		40	8,200
pnir7106/+5 p99G	IHF I	40	6,600	80

^aThe pnir7106 fragment carries pnir sequences from position −106 to +36. The pnir7106/+5 promoter fragment carries a 5-bp insertion at position −60 between the NarL/NarP and FNR binding sites (Fig. 1A). The p99G substitution disrupts IHF binding to IHF I (3).

^bβ-Gal activity of JCB3884 cells carrying pRW50 containing the indicated promoter fragment. Cells were grown aerobically (+ O₂) and anaerobically (−O₂) in minimal salts medium plus 0.4% glucose. Each activity is the average of three independent determinations that varied by less than 10%.

^cData represent the percentage difference in anaerobic expression due to the p99G substitution in IHF I.

Next, we used the pnir7133 fragment, together with the pnir7133/+5 and pnir7133/+10 derivatives with 5 and 10 bp, respectively, inserted between IHF I and IHF II at position −106 (Fig. 1A), to investigate the effects of IHF at IHF II. Results in Table 4 show that the 5-bp insertion decreased promoter activity, while the promoter with the 10-bp insertion gave similar expression levels to pnir7133. The p112G substitution, which disrupts IHF II, decreased expression from the pnir7133 and pnir7133/+10 promoters but not from the pnir7133/+5 promoter. Thus, IHF II must be correctly positioned on one face of the DNA helix for activation. To investigate the effects of IHF at IHF II in the absence of IHF I, we used the pnir7133 fragment carrying the p99G mutation. Results in Table 4 show that the activity of the pnir7133/p99G promoter was decreased about twofold by the p112G substitution. Thus, IHF II can stimulate pnir activity independently of IHF I.

TABLE 4.

β-Gal activities of JCB3884 cells carrying different pnir7133 fusions

Promoter fragmenta	Mutated site(s)	β-Gal activity (nmol of ONPG hydrolyzed min−1 mg−1 of dry cell mass)b		% Differencec
		+ O2	−O2
pnir7133		20	4,800
pnir7133/p112G	IHF II	20	2,100	44
pnir7133/+5		30	1,500
pnir7133/+5 p112G	IHF II	30	2,400	160
pnir7133/+10		30	4,200
pnir7133/+10 p112G	IHF II	30	1,800	43
pnir7133/p99G	IHF I	50	1,2600
pnir7133/p99Gp112G	IHF I and II	30	6,700	53

^aThe pnir7133 fragment carries sequences from position −133 to +36. The pnir7133/+5 and the pnir7133/+10 fragments carry 5- and 10-bp insertions, respectively, at position −106 between IHF I and IHF II (Fig. 1A). The p99G and p112G substitutions disrupt IHF I and IHF II respectively (3, 5).

^bβ-Gal activity of JCB3884 cells carrying pRW50 containing the indicated promoter fragment. Cells were grown aerobically (+ O₂) and anaerobically (−O₂) in minimal salts medium plus 0.4% glucose. Each activity is the average of three independent determinations that varied by less than 10%.

^cData represent the percentage difference in anaerobic expression due to the p112G substitution within IHF II.

To study the effects of Fis, we used the pnir7150 fragment together with the pnir7150/+5 and pnir7150/+10 derivatives in which 5 and 10 bp, respectively, were inserted between IHF II and Fis I at position −134 (Fig. 1A). Results in Table 5 show that both the 5- and 10-bp insertions resulted in increased promoter activity. The p146A substitution, which disrupts Fis I, increased expression from the pnir7150 and pnir7150/+10 promoters but not from the pnir7150/+5 promoter. From this we conclude that Fis must be correctly positioned on one face of the DNA helix for repression. To investigate the effects of Fis at Fis I in the absence of IHF I or IHF II, we used the pnir7150 fragment carrying either the p99G or the p112G mutation. Results in Table 5 show that the activity of the pnir7150/p99G promoter and the pnir7150/p112G promoter was increased by the p146A substitution. Similarly, with the pnir7150 promoter carrying both p99G and p112G, expression was increased by the p146A substitution. Thus, Fis can repress transcription in the absence of IHF at IHF I and IHF II. However, in each case tested, Fis-dependent repression is less efficient when one or both of the DNA sites for IHF is inactivated.

TABLE 5.

β-Gal activities of JCB3884 cells carrying different pnir7150 fusions

Promotera	Mutated site(s)	β-Gal activity (nmol of ONPG hydrolyzed min−1 mg−1 of dry cell mass)b		% Differencec
		+ O2	−O2
pnir7150		30	1,600
pnir7150/p146A	Fis I	30	3,700	230
pnir7150/+5		40	6,600
pnir7150/+5 p146A	Fis I	30	6,300	95
pnir7150/+10		30	3,400
pnir7150/+10 p146A	Fis I	30	6,500	191
pnir7150/p99G	IHF I	30	8,100
pnir7150/p99Gp146A	Fis I, IHF I	60	12,700	157
pnir7150/p112G	IHF II	30	700
pnir7150/p112Gp146A	Fis I, IHF II	30	1,200	171
pnir7150/p99Gp112G	IHF I, IHF II	40	3,700
pnir7150/p99Gp112Gp146A	Fis I, IHF I, IHF II	40	6,200	168

^aThe pnir7150 fragment carries sequences from position −150 to +36. The p99G, p112G, and p146A substitutions disrupt the IHF I, IHF II, and Fis I binding sites, respectively (3, 5, 28). In the pnir7150/+5 and pnir7150/+10 promoter fragments, insertions of 5 or 10 bp, respectively, occur at position −134 between Fis I and IHF II (Fig. 1A).

^bβ-Gal activity of JCB3884 cells carrying pRW50 containing the indicated promoter fragment. Cells were grown aerobically (+ O₂) and anaerobically (−O₂) in minimal salts medium plus 0.4% glucose. Each activity is the average of three independent determinations that varied by less than 10%.

^cData represent the percentage difference in anaerobic expression due to the p146A substitution in Fis I.

IHF and Fis repress open complex formation at pnir in vitro.

The effects of IHF and Fis on FNR-dependent activation at pnir were investigated using potassium permanganate footprinting to monitor open complex formation on an end-labeled pnir7150 fragment. In these experiments, we used purified RNA polymerase holoenzyme and purified FNR carrying the DA154 substitution that renders FNR active under aerobic conditions (27). Recall that permanganate cleaves Ts within single-stranded DNA, enabling the detection of DNA unwinding at promoters. The results, illustrated in Fig. 7, show that with RNA polymerase alone, unwinding at pnir was not detected, but with FNR present, open complex formation was clearly observed (Fig. 7A, lane 2). When either IHF or Fis was included, open complex formation at pnir was repressed (lanes 3 and 4 and lanes 6 and 7, respectively). With both Fis and IHF present, open complex formation was repressed further in comparison to each protein alone (Fig. 7B, compare lanes 3 and 6 with lane 7). Quantification of promoter opening, using the permanganate sensitivity at position −7, indicates that Fis and IHF work synergistically to repress pnir (Fig. 7B).

FIG. 7.

IHF and Fis repress FNR-dependent promoter opening at pnir in vitro. The figure shows in vitro potassium permanganate footprint experiments with purified FNR DA154, IHF, and Fis proteins. (A) The end-labeled pnir7150 AatII-HindIII fragment was incubated with RNA polymerase (RNAP), FNR DA154, IHF, and Fis and subjected to potassium permanganate footprinting. The concentration of FNR was as follows: lane 1, no protein; lanes 2 to 7, 250 nM. The concentration of IHF in each reaction mixture was as follows: lanes 1, 2, and 5 to 7, no protein; lane 3, 465 nM; lane 4, 930 nM. The concentration of Fis in each reaction mixture was as follows: lanes 1 to 4, no protein; lane 5, 447 nM; lane 6, 894 nM; lane 7, 1.79 μM. (B) End-labeled pnir7150 AatII-HindIII fragment was incubated with RNA polymerase (RNAP), FNR DA154, IHF, and Fis and subjected to potassium permanganate footprinting. The concentration of FNR was as follows: lane 1, no protein; lanes 2 to 9, 250 nM. The concentration of IHF in each reaction mixture was as follows: lanes 1, 2, and 6, no protein; lanes 3 and 7, 233 nM; lanes 4 and 8, 465 nM; lanes 5 and 9, 930 nM. The concentration of Fis in each reaction mixture was as follows: lanes 1 to 5, no protein; lanes 6 to 9, 894 nM. All lanes contained 50 nM RNA polymerase. Gels were calibrated using Maxam-Gilbert G+A (lane GA) sequencing reactions, and the location of cleavage sites produced by potassium permanganate footprinting within pnir are shown. Promoter unwinding in panel B was quantified using the permanganate cleavage at position −7, and values are given as a percentage of the cleavage observed in the presence of FNR only (lane 2).

Gel retardation assays were also used to examine the binding of RNA polymerase, FNR, IHF, and Fis to the labeled pnir7150 promoter fragment. Results shown in Fig. 8 indicate that when purified RNA polymerase holoenzyme was incubated with FNR, a stable ternary complex was formed (Fig. 8A and B, lanes 5). However, this complex disappeared when either IHF or Fis was present (Fig. 8A and B, lanes 7), indicating that both IHF and Fis repress FNR-dependent transcription by interfering with RNA polymerase binding. The diffuse nature of retarded complexes in the presence of either IHF or Fis may suggest that the ternary complex dissociates during electrophoresis.

FIG. 8.

IHF and Fis interfere with the binding of RNA polymerase to pnir. The figure shows gel retardation assays with purified RNA polymerase, FNR DA154, IHF, and Fis proteins. (A) End-labeled pnir7150 EcoRI-HindIII fragment was incubated with RNA polymerase (RNAP), FNR DA154, and IHF. The concentration of RNA polymerase was as follows: lanes 1, 3, 4, 6, and 8, no protein; lanes 2, 5, and 7, 174 μM. The concentration of FNR in the reaction mixture was as follows: lanes 1, 2, 3, and 8, no protein; lanes 4 to 7, 1.5 μM. The concentration of IHF was as follows: lanes 1 to 5, no protein; lanes 6 to 8, 200 nM. (B) End-labeled pnir7150 EcoRI-HindIII fragment was incubated with RNA polymerase (RNAP), FNR DA154, and Fis. The concentration of RNA polymerase was as follows: lanes 1, 3, 4, 6 and 8, no protein; lanes 2, 5, and 7, 174 μM. The concentration of FNR was as follows: lanes 1 to 3 and 8, no protein; lanes 4 to 7, 1.5 μM. The concentration of Fis was as follows: lanes 1 to 5, no protein; lanes 6 to 8, 446 nM.

DISCUSSION

Induction of the E. coli nir operon is codependent on two environmental stimuli, the absence of oxygen and the presence of nitrate or nitrite ions, and these stimuli are signaled by FNR and NarL or NarP, respectively (25). However, in the absence of nitrate or nitrite ions, some nir operon expression can be measured, and our studies have shown that the nucleoid-associated proteins IHF and Fis play a key role in setting the level of this expression by modulating FNR-dependent activation of the nir promoter (3, 5, 28). Here, we demonstrate that upstream IHF and Fis binding sites must be correctly positioned, with IHF I and Fis I repressing FNR-dependent transcription and IHF II having a stimulatory role. Since DNA-bound IHF and Fis both bend their target sites sharply (by ∼140° and ∼90°, respectively), the upstream sequences at pnir are likely to be folded (21, 24). While there are many promoters where IHF and Fis function alone to regulate transcription, at the nir promoter they function in concert via an upstream sequence module that can be transplanted onto another FNR-dependent promoter (Fig. 6).

IHF and Fis repress FNR-dependent transcription at pnir by modulating the binding of RNA polymerase. Since the DNA sites for IHF and Fis are distal to the core promoter sequences, it is unlikely that their binding blocks direct access of RNA polymerase. We favor a model in which they prevent the RNA polymerase α subunit C-terminal domain from docking with upstream sequences, hence destabilizing polymerase binding. Although the binding of IHF and Fis to each of the upstream target sites can function independently, we were able to measure some weak synergy. For example, Fis I-mediated repression is more effective when upstream IHF sites are intact, suggesting that IHF may help position Fis for optimal repression.

Our studies demonstrate that the effects of nucleoid-associated proteins at promoters can be complex. It is especially intriguing that binding of IHF to IHF I and to IHF II at the nir promoter produces opposing effects, and it is the balance between these effects that sets the level of FNR-dependent expression in the absence of nitrate/nitrite-dependent activation via NarL or NarP. Although the overall organization of the nir promoter is conserved in other pathogenic E. coli strains and enteric bacteria, the balance between IHF I and IHF II varies. Hence, evolution appears to have adjusted the binding of IHF to the two targets in order to set the levels of anaerobic nir operon expression in the absence of NarL or NarP.