Promoter and regulon analysis of nitrogen assimilation factor, σ⁵⁴, reveal alternative strategy for E. coli MG1655 flagellar biosynthesis

Abstract

Bacteria core RNA polymerase (RNAP) must associate with a σ factor to recognize promoter sequences. Promoters recognized by the σ⁵⁴ (or σ^N) associated RNA polymerase are unique in having conserved positions around −24 and −12 nucleotides upstream from the transcriptional start site. Using DNA microarrays representing the entire Escherichia coli genome and promoter validation approaches, we identify 40 in vivo targets of σ⁵⁴, the nitrogen assimilation σ factor, and estimate that there are 70 σ⁵⁴ promoters in total. Immunoprecipitation assays have been performed to further evaluate the efficiency of our approaches. In addition, promoter consensus binding search and primer extension assay helped us to identify a new σ⁵⁴ promoter carried by insB-5 in the upstream of flhDC operon. The involvement of σ⁵⁴ in flagellar biosynthesis in sequenced E. coli strain MG1655 indicates a fluid gene regulation phenomenon carried by some mobile elements in bacteria genome.

INTRODUCTION

The upstream regulatory region of all bacterial genes or operons contains one or more promoter(s). This is a special DNA sequence that can be specifically recognized by the RNA polymerase sigma subunit to allow binding and initiation of transcription. A major mode of gene regulation occurs via the binding of sigma factors to these specific DNA sequences. Sigma factors are identified by their ability to bind to core RNA polymerase (RNAP) and by their ability to direct promoter-specific transcription.

The Escherichia coli housekeeping σ factor, σ⁷⁰, was the first prokaryotic σ factor to be purified and characterized (1). Since then, numerous sigma factors have been found and characterized in E. coli and other prokaryotic organisms (2–6). The seven known E. coli sigma factors (σ⁷⁰, σ⁵⁴, σ³², σ^S, σ^F, σ^E and σ^FecI) have been categorized into two families. The σ⁷⁰ family contains σ⁷⁰, σ³², σ^S, σ^F, σ^E and σ^fecI, whereas σ⁵⁴, because of differences in sequence, promoter architecture, and function, is placed in its own separate family (7,8). The intracellular levels of each individual σ factor change in response to growth transitions and environmental conditions (9,10) that play important roles in the regulation of gene expression.

σ⁵⁴ (σ^N) was identified as a sigma factor involved in the transcription of genes involved in the cellular assimilation of ammonia and glutamate under conditions of nitrogen limitation (11). σ⁵⁴ is structurally and functionally distinct from the other E. coli σ factors and shares very little if any sequence similarity with the primary σ factors. The three major differences that separate σ⁵⁴ from the σ⁷⁰ family of the other σ factors are: (i) unlike members of σ⁷⁰ family, σ⁵⁴ is able to bind promoter DNA in the absence of core RNA polymerase (7); (ii) regulatory proteins like NtrB and NtrC activate σ⁵⁴ holoenzyme (12,13); (iii) σ⁵⁴ recognizes promoter sequences with conserved GG and GC elements located −24 to −12 nucleotides upstream of the transcription start site (3,7). Although some bioinformatics approaches have been applied to search σ⁵⁴ consensus binding site in different bacteria species (14–17), no large-scale experimental effort has been undertaken to unravel in detail the σ⁵⁴ regulon in E. coli. Here, we present an updated list of σ⁵⁴-dependent promoters in E. coli. Computer programs, such as BioProspector and HMMer, have been utilized together to search and present the derivation of an extended consensus sequence for σ⁵⁴ binding. Different from previous computational methods that only focused on the upstream intergenic sequences extracted from the genomes of several bacterial species, we found 18% of σ⁵⁴-promoters are located within the coding region of known genes or between convergently transcribed genes. This suggests a previously uncharacterized regulatory function of σ⁵⁴. We also compare σ⁵⁴-dependent genes identified in this study with σ⁷⁰-dependent genes identified in a separate study (our unpublished data). We found that 14% of σ⁵⁴-dependent genes can be directly transcribed by σ⁷⁰ in vitro. This might indicate that bacteria use different promoter organizations to produce different regulatory outcomes in different environments. In addition, we also found a new σ⁵⁴-dependent promoter upstream of the flhDC operon in the sequenced strain MG1655 and provide an alternative explanation for the high motility of this sequenced strain compared with its closely related E. coli strains.

EXPERIMENTAL PROCEDURES

Reagents, strains and plasmids

All reagents were purchased from Sigma Chemical Company (St Louis, MO, USA) unless otherwise indicated. A 10X MOPS minimal media was prepared as described in Neidhardt et al. (18). The media were filter sterilized through a 0.2 µm filter and stored at 4°C. The defined media for log-phase cell growth contained 1× MOPS minimal media, 0.1% glucose, 0.66 mM K₂HPO₄. Neidhardt's; MOPS-based defined media are now available commercially from Teknova, Inc.

Because the E. coli Genechip probe set is based on the sequenced E. coli K-12 strain MG1655 (λ⁻ F⁻ ilvG⁻ rfb⁻50 rph-1, prototroph) (19), we chose this bacterial strain for use in our study. In order to disrupt the expression of σ⁵⁴ in E. coli, we used a simple and highly efficient method (20,21) to prepared in-frame deletion strains for σ⁵⁴ as described (22,23). For controllable induction of individual regulators in vivo, we used the P_Ltet promoter which is controlled by the repressor TetR to construct these overexpression vectors as described previously (23). A downstream gene can be induced in the presence of the inducer anhydrotetracycline (aTc). All strains used in this study were derivatives of E. coli K12 MG1655.

Growth conditions, preparation of cell lysates

All cultures were grown in a New Brunswick Gyrotory water bath shaker (model G76) with vigorous aeration unless otherwise indicated. For cultures of cells carrying antibiotic resistance markers, the media were supplemented with ampicillin (100 µg/ml), chloramphenicol (30 µg/ml), or kanamycin (50 µg/ml) where appropriate. For induction of σ⁵⁴ under the control of the anhydrotetracycline (aTc)-regulated promoter, aTc was added at a concentration of 100 ng/ml as described previously (22,23).

Escherichia coli MG1655 WT strain as well as derived deletion mutant strains were grown overnight in MOPS minimal media at 37°C in an air shaker with vigorous aeration (225 r.p.m.). Two microliters of the overnight culture was used to inoculate 100 ml of fresh MOPS minimal medium. When the culture density reached OD₆₀₀ 0.2, a 1000 µl portion of culture was harvested into a prechilled 1.5 ml Eppendoff tube and then immediately put on ice for 1 min before being centrifuged at 10 000g (12 000 r.p.m. for BECKMAN Microfuge^R) for 10 min at 4°C. The supernatant was removed and the cell pellet resuspended immediately in 40 µl lysis buffer (1× SDS) and heated at 75°C for 5 min to quickly lyse the cells and prevent changes in the intracellular levels of the sigma factors being measured. We confirmed the absence of σ⁵⁴ in the rpoN deletion strain by Western blot analysis using a monoclonal antibody (6RN3) (24).

Instead of using a σ³²/σ^F-inducible strain as shown in previous σ³²/σ^F regulon studies (22,23), we used strains carrying a plasmid with an aTc-inducible σ⁵⁴ gene in this work. The same experimental procedures for induction, collection and treatment of sample were performed as described below and in more detail in our σ³²/σ^F regulon papers (22,23).

RNA isolation, cDNA synthesis, labeling and hybridization for microarray experiments

For preparing the total RNA for microarray experiments, E. coli strains were grown overnight in MOPS minimal media at 37°C in an air shaker with vigorous aeration (225 r.p.m.). Two microliters of the overnight culture was used to inoculate 100 ml of fresh MOPS minimal medium. A total of 15 ml samples of culture (corresponding to 7.5 × 10⁹ cells) were taken for wild-type and mutant strains when the culture density OD₆₀₀ value reached 0.2 and the same amount of culture was taken before and 5 min after induction in σ⁵⁴-overexpression strains. RNA was stabilized immediately by mixing with a double volume of RNAprotect Bacterial Reagent (Qiagen) and incubated at room temperature for 10 min. Cells were centrifuged at 5800g for 20 min and cell pellets were stored at −80°C prior to RNA extraction.

Total nucleic acid was isolated using MasterPure kits (Epicentre) as described by the manufacturer. DNase I (Epicentre) was used to remove genomic DNA contamination. Total RNA was purified, precipitated and resuspended in diethylpyrocarbonate (DEPC)-treated water. The quality and integrity of the isolated RNA was checked by visualizing the 23S and 16S rRNA bands on a 2% agarose gel. A 10 µg of total RNA was mixed with 500 ng random hexamers and then was reverse transcribed for first strand cDNA by using the Superscript II system (Invitrogen). RNA was removed by using RNase H (Life Technologies) and RNase A (Epicentre). cDNA was purified by using Qiaquick PCR purification kit (Qiagen) and followed by partial DNase I digestion to fragment cDNA to an average length of 50–100 bp. The fragmented cDNA was 3′-end-labeled by using terminal transferase (New England Biolabs) and biotin-N6-ddATP (PerkinElmer) and was added to hybridization solution to load on Affymetrix GeneChip^R E. coli Antisense Genome Arrays. Hybridization was carried out at 45°C for 16 h. The arrays were then washed and subsequently stained with streptavidin, biotin-bound anti-streptavidin antibody and streptavidin-phycoerythrin (Molecular Probes) to enhance the signal. Arrays were scanned at 570 nm with 3 µm resolution using a confocal laser scanner.

Data analysis

Image analysis was carried out by Affymetrix® Microarry Suite 5.0 software. Cell intensity files were first generated from the image data files. An absolute expression analysis then computes the detection call, detection P-value and signal (background-subtracted and adjusted for noise) for each gene. Genes were considered up-regulated relative to the 0 time point (before induction)/wild-type strain sample if they had a 2-fold or greater increase in signal intensity and the signal intensity in the experiment had a log₂ value of at least 8.0 and a detect level equal one; the higher log₂ intensity values were used to limit the analysis to those genes for which we have a high degree of confidence in their level of expression.

Array design

The GeneChip® E. coli Antisense Genome Array was purchased from Affymetrix (catalog number: 900381). It contains in situ synthesized probe sets to detect the antisense strand of more than 4200 known open reading frames and over 1350 intergenic regions. A given gene is represented by 15 different 25-mer oligonucleotides that are designed to be complementary to the target sequence (25–27). Sequence information for probes on the array corresponds to the M54 version of the E. coli Genome Project database at the University of Wisconsin. Complete array information, including the location for each feature on the array, can be found at www.affymetrix.com.

Purification and fluorescence labeling of proteins and MAbs for immunoblot assay

Purified core RNA polymerase was made from E. coli MG1655 according to the method of Thompson et al. (28). Purified sigma factors and monoclonal antibodies (MAbs) were made as described in Anthony et al. (24). Purified core RNA polymerase and sigma factors were used in in vitro transcription assays. Mouse MAbs used in this experiment were anti-β′ (NT73) and anti-σ⁵⁴ (6RN3) for measuring the intracellular level changes of σ⁵⁴. Both MAbs are available from Neoclone (Madison, WI, USA). Fluorescent dye, IC5-OSu (Dojindo), was used to label the primary antibodies according to previously described methods (29). The IC5-labeled MAbs, at final stored concentrations of 1 mg/ml, were diluted 1: 2000 for use in this experiment. Electrophoresis and immunoblot assays were performed as described in a previous paper (22,23). Signal intensities of the bands obtained with the Molecular Dynamics Typhoon system were quantified using the ImageQuant program.

Chromatin immunoprecipitation assays

Chromatin immunoprecipitation (ChIP) assay is usually used to crosslink proteins to adjacent DNA by adding formaldehyde to an in vivo culture. Recently, the Aseem Ansari group at UW-Madison has developed the current ChIP protocol (30,31) for E. coli studies. To map promoters in bacteria, they sought a way to force RNAP to reside only at promoters so that identifying DNA fragments bound to a given sigma factor associated RNAP in vivo would report promoter locations. A variety of small-molecule inhibitors of RNA polymerase were evaluated for the immobilization of RNAP, and rifampin was found to work best (31). The antibiotic rifampin inhibits bacterial growth by binding the ß-subunit of RNAP near the active site, blocking the synthesis of RNAs longer than 2–3 nt (32). Rifampin has no effect on RNAP promoter binding (33) and has no effect on RNAP in vitro when added after elongating RNAP has cleared the promoter (34,35).

The general procedures for the confirmation of σ⁵⁴-dependent genes are: cross-linked chromatin is isolated from rifampin-treated σ⁵⁴ overexpression and deletion cells, cells were then resuspended in 500 ml lysis buffer [10 mM Tris–HCl (pH 8.0), 50 mM NaCl] with 3 mg/ml lysozyme and incubated for 30 min at 37°C. A 500 ml of 2 × IP buffer [200 mM Tris–HCl (pH 8.0), 600 mM NaCl, 4% Triton X-100] was added and the DNA was sheared by sonication. Cell debris was removed by centrifugation, and the supernatant was transferred to new tubes. The samples were immunoprecipitated with monoclonal antibodies against σ⁵⁴ (6RN3) (NeoClone, Madison, WI, USA). The immunoprecipitated protein-DNA crosslinks are then reversed at 65°C overnight. The ChIP DNA served as template for amplifying σ⁵⁴-dependent promoters by PCR.

Primer extension assays

The total RNA was prepared as described in microarray experiments above. The primer (5′-GTTGCGATAAGCTGCAA) was 5′-end-labeled using T4 polynucleotide kinase (New England Biolabs) and 50 μCi of [γ-³²P]-ATP (PerkinElmer). Approximately 1.5 pmol of ³²P-end-labeled oligonucleotide was added to the reaction. To denature nucleic acids, the reactions were heated at 95°C for 3 min and quenched on wet ice. An incubation step at 50°C for 15 min was done to promote annealing of the oligonucleotide to the RNA template. Two units of AMV Reverse Transcriptase (Promega) and final concentrations of 1 × AMV Reverse Transcriptase buffer [50 mM Tris–HCl (pH 8.3), 50 mM KCl, 10 mM MgCl₂, 0.5 mM spermidine, 10 mM dithiothreitol (Promega)], and 0.2 mM each of the four deoxynucleoside triphosphates were added to each reaction tube. Reaction mixtures were incubated at 42°C for 90 min. Primer extension products were electrophoresed on a 5% denaturing polyacrylamide gel. The gel was dried, exposed in a phosphorimaging cassette, and scanned by using a Molecular Dynamics Typhoon (Model 8600).

RESULTS

The level of σ⁵⁴ in E. coli MG1655 derivative deletion and overexpression strains

Escherichia coli σ⁵⁴ in-frame deletion strains as well as σ⁵⁴-overexpression strains were constructed as described in Zhao et al. (22,23) and in Experimental procedures section. A fast and reliable improved western blot assay was used for quantitative analysis of the intracellular level of σ⁵⁴ in the in-frame deletion and overexpression strains. The β′-subunit of core RNA polymerase was also examined to serve as an internal control because its intracellular levels remain constant under various conditions (9,36). The signal intensities of the proteins were immunodetected by corresponding IC5-labeled monoclonal antibodies. Our results (Figure 1A) show that the σ⁵⁴ protein was not expressed (Figure 1A) in the rpoN deletion strain, confirming inactivation of this gene. The σ⁵⁴ protein level, which is normalized to the β′ subunit of RNA polymerase, rapidly increased 5 min after induction with an ∼2.9-fold change. Using the same overexpression system, previous data (22,23) showed σ³² and σ^F protein levels after 5 minutes induction increased almost 7.4- and 2.3-fold, respectively. The discrepancy of the fold changes, for different σ factors under the same P_Ltet promoter control, is mainly due to the fact that the experiments were performed at log-phase (OD₆₀₀ = 0.2) in minimum medium, in which different σ factors have different initial protein levels.

Figure 1.

Determination the σ⁵⁴ protein level in the rpoN deletion (KZ30) and σ⁵⁴ overexpression (KZ7) strains, respectively. (A) Western blot analysis of β′ and σ⁵⁴ protein expression in the wild-type MG1655 as well as in the rpoN deletion strains. Expression of β′ subunit of core RNA polymerase, which served as internal controls, can be detected in both strains. Expression of σ⁵⁴ can only be detected in wild-type strain, but the expression of rpoN gene is absent in the respective mutant strain. (B) Left: western blot of β′ and σ⁵⁴ expression before and 5 min after induction. Right: quantification of western blot. The σ⁵⁴ protein level increases ∼2.9-fold after 5 min of induction. Signal intensities are determined using ImageQuant version 5.2 software.

Known σ⁵⁴-dependent promoters

To characterize the effect of the decreasing and increasing σ⁵⁴ protein level in vivo on gene expression, global RNA transcript abundance was monitored in the deletion mutant strain and the overexpression strain 5 min after σ⁵⁴ induction with cells grown in log-phase (OD₆₀₀ = 0.2) in MOPS minimal medium at 37°C. Transcription profiles were obtained as described in ‘Experimental procedures’ section. The sample for the wild-type strain and for the overexpression strain at time zero before induction was used as the reference to identify genes whose transcript abundance had significantly changed in the rpoN deletion mutant strain or the induced σ⁵⁴-overexpression strain, respectively.

DNA microarray results showed the transcriptional level of all well-characterized genes belonging to the σ⁵⁴ regulon are downregulated/upregulated in the rpoN deletion strain/σ⁵⁴ overexpression strain (Table 1). These results are consistent with our previous hypothesis that a change of the intracellular level of a given sigma factor will cause a change of the transcriptional level of genes dependent on this sigma factor. Jishage (9) reported that the intracellular level of σ⁵⁴ is maintained at 16% or 6% the level of σ⁷⁰ during log and stationary phase growth, respectively, in two different strains. Loss of σ⁵⁴ in cells will decrease the transcription of σ⁵⁴-dependent genes. Induction of σ⁵⁴ will show an increase in the transcriptional level of σ⁵⁴-dependent genes.

Table 1.

18 Known σ⁵⁴-dependent genes with their promoter sequence

b#a	Geneb	Product	Function	Overc	Deld	Starte	End	promoter_sequencef
b0450	glnKg	regulatory protein (P-II 2) for nitrogen assimilation, regulates GlnL (NRII), GlnE (ATase), and AmtB (ammonium transporter)	Central intermediary metabolism	1.5	−18.8	−68	−51	tttcTGGCACACCGCTTGCAATacct
b3870	glnAg	glutamine synthetase	Amino acid biosynthesis: Glutamine	1.2	−11.4	−99	−82	aagtTGGCACAGATTTCGCTTTatct
b0811	glnHg	high-affinity glutamine transport protein	Transport of small molecules	1.1	−8.3	−70	−53	aaacTGGCACGATTTTTTCATAtatg
b1988	nac	transcriptional repressor of histidine utilization/nitrogen assimilation (LysR family)	Central intermediary metabolism	1.3	−4.2	−71	−54	aaacTGGCAAGCATCTTGCAATctgg
b1304	pspAg	phage shock protein; negative regulatory gene for the psp operon	Phage-related functions and prophages	2.1	−3.9	−67	−50	aaatTGGCACGCAAATTGTATTaaca
b3227	dcuD	putative C4-dicarboxylate transport protein (DcuC family)	Not classified	12.5	−3.3	−313	−296	aaaaTGGCAGGGTTTTCTCTTTgtgc
b4002	zraP	periplasmic Zn-binding protein, zinc resistance-associated	Unknown	4.1	−3.1	−51	−34	tcgtTGGCACGGAAGATGCAATaccc
b0331	prpB	putative carboxyphosphonoenolpyruvate mutase	Not classified	2.2	−2.8	−62	−45	attgTGGCACACCCCTTGCTTTgtct
b4003	zraS	sensory histidine kinase in two-component regulatory system with ZraR, regulates zraP expression, senses Zn	Energy metabolism, carbon: Fermentation	3.5	−2.3	−47	−30	aagaTGGCATGATTTCTGCTGTcaga
b2725	hycAg	regulatory protein for HycE (part of the FHL complex)	Energy metabolism, carbon: Fermentation	13.0	−2.3	−52	−35	aagtTGGCACAAAAAATGCTTAaagc
b2221	atoDg	acetyl-CoA:acetoacetyl-CoA transferase, alpha subunit	Degradation of small molecules: Fatty acids	12.6	−2.3	−61	−44	attcTGGCACTCCCCTTGCTATtgcc
b4079	fdhF	formate dehydrogenase H, selenopolypeptide subunit	Energy metabolism, carbon: Anaerobic respiration	3.0	−2.2	−67	−50	aatgTGGCATAAAAGATGCATActgt
b3686	ibpBg	small heat shock protein	Adaptations, atypical conditions	4.4	−2.1	−185	−168	aaccTGGTAAATGGTTTGCTGTatat
b1748	astCg	succinylornithine transaminase, also has acetylornitine transaminase activity, PLP-dependent	Amino acid biosynthesis: Arginine	3.1	−2.1	−88	−71	tggcTGGCACGAACCCTGCAATctac
b3421	rtcBg	conserved hypothetical protein	Unknown	2.1	−2.1	−54	−37	tttcTGGCACGACGGTTGCAATtatc
b2713	hydNg	electron transport protein (formate to hydrogen), Fe-S center	Energy metabolism, carbon: Anaerobic respiration	1.4	−1.7	−53	−36	aaacTGGCATGATTTGTGAATGtatc
b2726	hypAg	guanine-nucleotide-binding protein in formate-hydrogenlyase system, functions as nickel donor for HycE of hydrogenlyase 3	Energy metabolism, carbon: Anaerobic respiration	2.4	−1.3	−46	−29	acacTGGCACAATTATTGCTTGtagc
b3073	ygjG	putative acetylornithine aminotransferase, PLP-dependent	Amino acid biosynthesis: Arginine	7.1	−1.2	−35	−18	ggagTGGCGCAATCCCTGCAATactt

^ab no. indicates Blattner number.

^bIt is possible that one gene has several different gene names.

^cNumbers indicate fold increase relative to pre-σ^N induction.

^dNumbers indicate fold decrease relative to wild-type strain.

^eNumbers indicate the distance from the gene’s translation start site.

^fPotential σ^N-related promoter

^gThe first gene in a known or predicted multicistronic operon.

There are 18 known promoters (controlling 52 genes) under the control of σ⁵⁴ in E. coli. Using a stringent cut off (2-fold decrease/increase to reduce the potential noise caused by array signal variation), 15 out of these 18 promoters significantly downregulated its operon genes in the rpoN deletion strain and 13 out of these 18 significantly upregulated their controlled genes in the σ⁵⁴ overexpression strain. This indicates our microarray experiments can detect most known σ⁵⁴-dependent promoters in our assays.

New candidate genes for σ⁵⁴ regulon

Expression profiling of transcripts corresponding to the complete set of ORFs in the E. coli genome revealed that the response to the changes of σ⁵⁴ in vivo was quite broad. In addition to identifying the known σ⁵⁴-dependent genes, our microarray data allowed us to assign many additional new candidate genes to the σ⁵⁴ regulon. Comparative analysis of the microarray data from the set of genes whose transcription is downregulated in the rpoN deletion strain (decrease of σ⁵⁴) and the set of genes with increased transcription at 5 minutes after σ⁵⁴ induction (increase of σ⁵⁴) allows us to narrow down to 22 new candidate genes with high confidence in the σ⁵⁴ regulon (Table 2). Results from promoter region consensus analysis using the algorithms MEME (37) and BioProspector (38) revealed the upstream regulatory sequences of most newly identified genes have a good match with the previously known σ⁵⁴ consensus binding site (Table 2).

Table 2.

22 New genes for the σ⁵⁴ regulon

b#a	Geneb	Product	Function	Overc	Deld	Starte	End	promoter_sequencef
b0319	yahEg	conserved protein	Unknown	2.6	−5.4	−215	−198	ctacTGGAAGCGATTGTGCTTAatga
b0045	yaaU	putative transport protein (MFS family)	Not classified	5.7	−4.9	−244	−227	aaacAGGCGCTGGAGCTGCTGGtgaa
b0534	sfmF	putative fimbrial-like protein	Not classified	3.2	−3.4	−347	−330	ggccGGGTAATCGACCTGCTGGtgtc
b1337	abgBg	putative amidohydrolase (aminoacylase), p-aminobenzoyl-glutamate utilization	Unknown	7.5	−3.1	−169	−152	atgaTGGCCCGCGTGCAGCAACatca
b0240	crl	transcriptional regulator of cryptic genes for curli formation and fibronectin binding	Surface structures	2	−2.9	−34	−17	aattTGGTAAAACAGTTGCATCacaa
b3476	nikAg	nickel transport protein (ABC superfamily, peri_bind)	Transport of small molecules: Cations	3.3	−2.9	−257	−240	cgccTGGCAAATCGTCAGCGTAgaca
b3673	emrD	multidrug transport protein (MFS family)	Drug/analog sensitivity	2.2	−2.7	−100	−83	ttccTGGCGTATATCTGGCTAAcatt
b1012	ycdMg	putative enzyme	Unknown	2.6	−2.7	−44	−27	aaacTGGCATCCGCTTTGCAAAcaag
b1296	ycjJ	putative transport protein (APC family)	Not classified	5.8	−2.6	−161	−144	gttaTGGAGCGCGGGCGGCAACgggc
b3521	yhjC	putative transcriptional regulator (LysR family)	Not classified	2.5	−2.5	−234	−217	ttggTGGTTAGTACGCATGCAATtaa
b0364	yaiS	conserved hypothetical protein	Unknown	7.6	−2.5	−49	−32	gctgTGGCGCATCGCTTGCTCGtctt
b2878	ygfKg	“putative oxidoreductase, Fe-S subunit”	Not classified	2.1	−2.4	−97	−80	aaccTGGCAAGAGTGGTGCGATtgtt
b3800	aslB	putative transcriptional activator of acrylsulfatase synthesis	Not classified	3.4	−2.4	−206	−189	agctTGGTAGCGCAACTGGTTTggga
b2710	b2710g	flavorubredoxin (FIRd) with NO-binding non-heme diiron center	Not classified	9.8	−2.3	−63	−46	aaacTGGCACGCAATCTGCAATtagc
b3383	yhfZ	unknown CDS	Unknown	2.6	−2.3	−91	−74	ccgtTGGCCTGACGCAGGCCGCgttg
b4067	yjcG	putative transport protein (SSS family)	Not classified	2.1	−2.2	−62	−45	catcTGGCGGGCGAACGGCGAAttcg
b2184	yejH	putative ATP-dependent helicase	Not classified	2.8	−2.1	−39	−22	tccaTGGCATACTATTAGCAGAataa
b1488	ddpXg	D-Ala-D-Ala dipeptidase, Zn-dependent	Unknown	2.7	−2.1	−60	−43	ggcaTGGCATGAGATCTGCATAagcg
b0473	htpG	chaperone Hsp90, heat shock protein C 62.5	Chaperones	3.7	−2.1	333	316	cgtcTGGAACAGCGTCTGGCAGAggaa
b2866	xdhAg	putative xanthine dehydrogenase subunit, molybdenum cofactor-binding domain	Unknown	4.9	−2.1	−110	−93	tttcTGGCGTAAATCTTGCCTGctta
b3902	rhaD	rhamnulose-1-phosphate aldolase	Degradation of small molecules	3.6	−2	−143	−126	tatcAGGCCTACAGGTCGGCAATagtt
b2470	acrD	aminoglycoside/multidrug efflux pump (RND family)	Drug/analog sensitivity	2	−2	−74	−57	cgatTGGCTCGTACCTTGCCGCtaca

^ab no. indicates Blattner number.

^bIt is possible that one gene has several different gene names.

^cNumbers indicate fold increase relative to pre-σ^N induction.

^dNumbers indicate fold decrease relative to wild-type strain.

^eNumbers indicate the distance from the gene’s translation start site.

^fPotential σ^N-related promoter

^gThe first gene in a known or predicted multicistronic operon except yahE and abgB.

To further confirm new genes in the σ⁵⁴ regulon, we performed ChIP assays to test the binding of σ⁵⁴-associated holoenzyme to promoter regions of these genes in vivo. We used the ibpB gene as positive control for the ChIP assay because this gene is known to be under the control of σ⁵⁴ (39). The upstream sequence of the dnaK gene was chosen as a negative control because transcription of this gene is regulated by σ⁷⁰ and σ³² but is not σ⁵⁴-dependent (40).

Using specific monoclonal antibodies (mAb) against σ⁵⁴, immunoprecipitation (ChIP) assays show the ibpB gene’s promoter DNA sequence can be pulled down by anti-σ⁵⁴ mAb in σ⁵⁴ overexpression strain, but not in σ⁵⁴ deletion strain due to lack of functional σ⁵⁴. For the negative control, the promoter DNA sequence of dnaK cannot be pulled down in either the σ⁵⁴ overexpression or the σ⁵⁴ deletion strain by anti-σ⁵⁴ mAb in vivo because the gene lacks a σ⁵⁴ promoter in its regulatory region. In addition, our ChIP assays show that 8 out of 10 of the newly identified genes’ promoter DNA fragments (as shown in Figure 2) can be directly pulled-down in the σ⁵⁴-overexpression strain and not in the σ⁵⁴-deletion strain. This indicates that most genes identified by our microarray experiments are genuine.

Figure 2.

Chromatin immunoprecipitation assays to test the DNA fragments carrying putative σ⁵⁴ promoter element(s). (A) ibpB served as a positive control and dnaK is serve as negative control. ChIPs assays are performed by using specific monoclonal antibody against σ⁵⁴ (6RN3). The experiments are performed in both σ⁵⁴ overexpression (left) and σ⁵⁴ deletion (right) strains, respectively. The samples of DNA fragments containing σ⁵⁴-dependent promoters were immunoprecipitated by monoclonal antibodies against σ⁵⁴. (B) Potential σ⁵⁴ consensus binding sites of each gene are predicted and aligned by computer program. The previously known σ⁵⁴ two-block promoter element consensus is shown below.

In a separate study, we identified σ⁷⁰ targets comprehensively across the E. coli genome. We compare those σ⁷⁰ target genes with the σ⁵⁴-dependent genes identified in this study. We found 14% of the σ⁵⁴-dependent genes can be transcribed by σ⁷⁰-associated RNAP (unpublished data), indicating bacteria use different promoter organizations to produce different regulatory outcomes under the appropriate environment conditions (41).

Computer prediction of σ⁵⁴-related promoter elements

A computer program was used to examine upstream DNA sequence of upregulated genes in our microarray data to look for regulatory sequence motifs. As prokaryotic promoter motifs often occur in two blocks with a gap of variable length, BioProspector (38), a C program which is capable of modeling motifs with two blocks and uses a Gibbs sampling strategy, was used to find the −12 and −24 consensus regions for σ⁵⁴ binding. Upstream sequences (400 bases from the first genes in transcription units that contain 2-fold up-regulated/down-regulated genes in our microarray data) were extracted as input sequences. A number of overall highest scoring motifs as position- specific probability matrices were reported. According to the reported highest scoring motif and its site locations on the input sequence, a graphical display of the results was generated using SEQUENCE LOGO (42) (Figure 3). The resulting consensus is represented as a TGGca-(N)_(4–5)-ttGCaa, where lower case indicates a less highly conserved site. This consensus agrees well with previously reported Eσ⁵⁴ consensus which was aligned to maximize alignment (TGGcacg-(N)_4–5-tGCtat) in the −24 and −12 regions of several published Eσ⁵⁴ promoters (3,43–45), indicating the conservation of σ⁵⁴ promoter across these genes.

Figure 3.

Determination of the σ⁵⁴ consensus binding site. σ⁵⁴-related two-block promoter element is aligned using Bioprospector (38) from the upstream sequence of genes in σ⁵⁴ regulon identified in our assays and displayed using SEQUENCE LOGO (42). The height of each column reflects the non-random bias of particular residues at that position, the size of each residue letter reflecting its frequency at that position.

An alternative strategy for E. coli MG1655 flagellar synthesis

In Gram-negative bacteria, the hierarchy of the flagellar regulatory system has been well characterized in micro-organisms with peritrichous flagella, such as E. coli and Salmonella typhimurium (46,47) and polar flagellated micro-organisms, such as Caulobacter crescentus (48). Recently, the regulatory cascade components of bacteria with one polar flagellum, such as Vibrio cholerae, Pseudomonas aeruginosa and P. fluorescens, have been characterized (49–51). The regulators at the top of the polar flagellar hierarchy belong to the NtrC family of σ⁵⁴-associated transcription activators. This regulator, together with σ⁵⁴, activates the expression of genes in polar flagellar system. In addition to its role in flagellar gene expression in these bacteria, this alternative sigma factor is known to participate in transcription of genes in nitrogen assimilation in E. coli and S. typhimurium (11,52) and in pilin synthesis in P. aeruginosa and Neisseria gonorrhoeae (53). No experiments have been performed and reported to test whether or not σ⁵⁴ is involved in E. coli flagellar biosynthesis.

To investigate this hypothesis, we first compared the motility of wild-type strain and rpoN in-frame deletion strain by growing them on the swarm plate. We found the rpoN null mutant significantly reduces its motility compared with its derivative wild-type strain (Figure 4). Our microarray data also showed a significantly down-regulation of the transcription of flagellar genes in rpoN in-frame deletion strain (Supplementary Table S2). Because the expression of the FlhDC operon can be regulated by multiple positive and negative regulators (Supplementary Figure S1) (54–57), we checked whether σ⁵⁴ affects the expression of these regulators and thus indirectly affects flhDC operon expression. Our microarray data show the expression of most negative regulators is downregulated and the expression of most positive regulators is upregulated (Supplementary Table S3) in σ⁵⁴ deletion strain. Therefore, the downshifting of expression of the flhDC gene in the σ⁵⁴ deletion strain is not primarily due to the level changes of those regulators in the same strain.

Figure 4.

Motility in E. coli wild-type and the rpoN deletion strains. Compared with wild-type strain, disruption of rpoN causes impaired movement on a swarm plate. The motility can be complemented from this mutant strain by in vivo expression of σ⁵⁴ from a plasmid-borne rpoN gene.

The Matsumura group found the sequenced E. coli strain MG1655 is extremely motile compared with other E. coli strains (58). Sequence analysis shows that there are mobile elements, insA-5 and insB-5, inserted into the upstream regulatory region of flhDC operon. They proposed that the high motility of MG1655 is due to insertion element insB-5, which prevents binding of the negative regulator OmpR, and thus increase the expression of flhDC operon for flagellar synthesis (Supplementary Figure S1). We think we can provide an alternative explanation for at least some of this high motility phenomenon; a σ⁵⁴-dependent promoter has been brought in by this insertion element.

A bioinformatics approach (38) has been used to search the σ⁵⁴ consensus binding site in the upstream regulatory region of flhDC operon. A good match of a σ⁵⁴–dependent promoter has been found in this insB-5 insertion element (Figure 5A). RT-PCR as described in previous paper (22,23) (PrimerUp: 5′-GCATGACAAAGTCATCGG, PrimerDown: 5′-GTTGCGATAAGCTGCAA) has been performed and showed that there is an additional transcript from the upstream region of the known σ⁷⁰-dependent promoter on this operon (Figure 5B). Primer extension assays showed this transcript present in the σ⁵⁴-overexpression strain but not in the σ⁵⁴-deletion strain (as shown in Figure 5C), indicating it was transcribed from a σ⁵⁴-dependent promoter that was brought in by the insB-5 insertion element. The biological significance of this is discussed below.

Figure 5.

Evidence indicating potential new σ⁵⁴ promoter in the upstream of flhDC operon. (A) Schematic show the positions of insertion elements (insA-5 and insB-5) and the σ⁷⁰, σ⁵⁴ promoters in the upstream of flhDC operon. Two-block binding site for these two promoters have been selected to compare with respective known consensus promoter binding sites as shown below. (B) Total RNA was prepared and reverse transcribed to cDNA. The primer pair as shown in (A) was used to amplify the transcripts starting from upstream of the σ⁷⁰-dependent promoter in the regulative region of the flhDC operon. (C) Primer extension assays have been performed in the σ⁵⁴-overxpression strain and deletion strain. A σ⁵⁴-dependent transcript is present in the σ⁵⁴ overexpression strain and is not in the σ⁵⁴ deletion strain. The σ⁷⁰-dependent transcript can be detected in both lanes.

DISCUSSION

Transcription is a key control point for regulation of numerous cellular activities. Bacteria regulate levels of gene expression by using transcription factors that modulate the recruitment of RNAP to promoter elements in the DNA. Because σ factors are required to initiate gene transcription in E. coli, and there are different σ factors required to regulate different sets of genes or regulons to adapt to changes in external environment, determination of σ regulons will provide valuable information for classifying genes into different functional groups and will be the first essential step to understanding global gene regulation under different growth conditions.

The aim of the present work is to identify promoters under control of Eσ⁵⁴ in vivo. Under defined, steady-state growth conditions, we used two different genetic approaches to alter σ⁵⁴ concentration in cells: (i) moderately expressing σ⁵⁴ from a plasmid-borne rpoN gene controlled by anhydrotetracycline (aTc)-inducible and Tet repressor-controlled P_Ltet promoter; (ii) disrupting the expression of σ⁵⁴ in rpoN deletion mutant strains. These combined methodologies used to determine Eσ⁵⁴ promoters proved highly effective as nearly all known σ⁵⁴-dependent promoters were identified. Our analysis has identified 40 targets of Eσ⁵⁴, including 22 previously undescribed targets. This is far more Eσ⁵⁴ promoters than expected. We chose a stringent cutoff (less than 2-fold change) when analyzing the microarray data to ensure that almost all called targets are genuine targets of Eσ⁵⁴. Our site validation and ChIP assays confirms that most, if not all, are genuine. We repeated our analysis using a less stringent cutoff (≥1.5-fold). In this case, we identified an additional 30 target regions (Supplementary Table S1) with strong σ⁵⁴-like promoter in its upstream regulation region. On the basis of this analysis, we estimate that there are 70 σ⁵⁴ promoters in E. coli, almost four times the number of previously identified promoters. Around 20% of the newly identified genes are hypothetical and the role of these genes remains to be elucidated. To obtain a better understanding of consensus binding sites controlled by σ⁵⁴, we carried out an in silico analysis on the presence of −24/−12 type promoters in the upstream region of these genes. Under the condition of our experiment, we think the expression of some of the identified σ⁵⁴-depenedent targets might be just a small fraction of the maximum due to strict dependence on AAA-family activators (59–62) some of whose activities might be limiting in our experiments. Likewise, it is possible that a few σ⁵⁴-dependent targets will be missed in our experiments due to very low or no activities of certain AAA-family activators.

We characterized an alternative strategy for sequenced E. coli strain MG1655 flagellar biosynthesis. In this case, σ⁵⁴-dependent promoter is brought by an insertion element into the upstream regulatory region of flhDC operon which encodes FlhDC, the master regulator for flagellar biosynthesis. Primer extension assays show there is a transcript transcribed from this new σ⁵⁴-dependent promoter in vivo. This also provides additional explanation for the high motility of this strain compared with other related bacteria. Lateral gene transfer is a major factor in the evolution of bacteria. Bacteria genomes contain a significant number of mobile elements—DNA that can move around on chromosomes, among organisms and even between species. It has been reported that mobile DNA can carry genes for virulence and drug resistance, as well as benign genes (63). The discovery of a σ⁵⁴-dependent promoter in a motile element in our work might provide another paradigm for the horizontal exchange of genetic information in prokaryotes. This additional fluidity allows bacteria to turn on different genes in various conditions to enhance their ability to survival or cause disease. The implication of σ⁵⁴ involvement in bacterial motility regulation is widespread in Gram-negative bacteria. Indeed, the elements of flagellar regulatory cascades including σ⁵⁴-associated regulators have also been identified in micro-organisms such as C. crescentus (48), Rhodobacter sphaeroides (64–66), Helicobacter pylori (67) and Campylobacter jejuni (68). Although environmental and genetic factors that control dissemination of these mobile elements remain to be determined, the σ⁵⁴ involvement in E. coli MG1655 flagellar biosynthesis might indicate a remarkable rearrangement/improvement in the functional organization of regulatory mechanisms that has existed in its other close-related bacteria. This might also indicate an evolutionary conservation among Gram-negative bacteria.

SUPPLEMENTARY DATA

Supplementary Data are available at NAR Online.

FUNDING

NIGMS (GM28575 to R.R.B.).

Conflict of interest statement. RRB acknowledges financial interest in NeoClone, LLC., (Madison, WI) which markets mAbs (NT73 and 6RN3) used in this study.