Introduction

Before orally ingested enteric pathogens such as Escherichia coli can reach their targeted mammalian host cells, they must first survive the extremely acidic conditions as they pass through the stomach (pH 2–4) on their way to the intestine. This is an extremely hostile environment, and thus E. coli contains multiple inducible systems for protection from acidic stress [11, 36].

Escherichia coli displays a high degree of acid resistance (AR) induction, classified as amino acid-dependent or amino acid-independent [9, 11, 18, 24]. The amino acid-independent system (AR1) is induced in cells grown to stationary phase in a moderately acidic medium, is suppressed by glucose, and is dependent on the RpoS sigma factor and the cyclic AMP receptor protein (Crp) [6]. The amino acid-dependent systems depend on the presence of specific amino acids. The glutamate-dependent system (AR2) requires two glutamate decarboxylases (GadA and GadB) plus GadC, an inner-membrane glutamate/γ-aminobutyrate antiporter [6, 15]. This is the most effective acid stress response pathway under extremely acid stress conditions, and over 20 proteins and 3 small noncoding RNAs have been identified in regulating the system [7, 14, 19, 28, 41, 47]. The complex regulation of AR2 is orchestrated by GadE, a global transcriptional activator [5, 23]. The AR3 system is arginine dependent and consists of the AdiA arginine decarboxylase and the AdiC arginine/agmatine antiporter [13]. It is induced by low pH and anaerobic conditions. AR4 is a lysine-dependent system that requires the CadA lysine decarboxylase and the CadB lysine/cadaverin antiporter [25, 43]. The expression of the cadBA operon is transcriptionally controlled by extracellular pH and lysine [30].

In addition to these structural enzymes and regulators, histone-like nucleoid-structuring proteins have also been reported to coordinate the AR systems in E. coli. H-NS was shown to regulate AR2 by inhibiting gadA and gadX transcription [12]. As a top-level regulator of AR with global influence, H-NS enhances degradation of rpoS mRNA [3] and represses expression of specific regulators of the amino acid-dependent AR systems [12, 19, 20]. More recently, another histone-like protein HU was shown to regulate expression of adiA and adiC, resulting in the induction of the arginine-dependent AR3 [2]. Hence, both H-NS and HU function as transcriptional regulators of AR induction.

In addition to H-NS and HU, E. coli possesses several other nucleoid architectural proteins that mimic the function of histones by assisting DNA folding and compaction. One of the best-characterized histone-like proteins is the integration host factor (IHF), a relatively abundant small DNA-bending protein [10]. IHF is a heterodimer which consists of α and β subunits, the products of the ihfA and ihfB genes [10, 29]. IHF, named for its role in phage λ integration [26], bends DNA by 160° [31] and facilitates the formation of nucleoprotein structures [42]. IHF binds to a specific DNA sequence and participates in a number of DNA activities including gene inversion, transposition, replication, partitioning of replicated DNA molecules, transfer of plasmids, and control of transcription [34, 39, 46].

IHF has been reported to act as a transcriptional regulator controlling the expressions of many specific operons [4, 8, 17]. IHF was also required for the induction of the arginine decarboxylase gene (adiA) expression at acidic pH [33]. This led us to examine the potential role of IHF in other AR structural gene expression and AR induction. In this report, we propose that IHF regulates the induction of amino acid-dependent AR systems via both transcriptional and translational controls of gene expression.

Materials and Methods

Bacterial Strains and Growth Conditions

The E. coli K-12 strains used in this study are listed in Table 1. The strains were grown at 37 °C in Luria–Bertani (LB) medium or minimal EG medium [16] consisting of E medium containing 0.15 mM uracil, 0.05 mM thymine, and 0.4 % glucose. The medium pH was adjusted with HCl or KOH. When required, sodium ampicillin (100 µg/ml), tetracycline (15 µg/ml), kanamycin (25 µg/ml), or chloramphenicol (20 µg/ml) were added. Bacterial growth was monitored by measuring absorbance at 600 nm. The cell optical density at an absorbance of 0.3 was approximately 2 × 108 colony-forming units (CFU)/ml under our experimental conditions.

Table 1 Bacterial strains, plasmids and oligonucleotide primers used in this study
Full size table

Acid Resistance (AR) Assay

AR was measured as described elsewhere [2]. Acid resistance is given as a percentage ratio of the number of surviving cells after acidic challenge at pH 2.5 to the number of cells before the challenge. When indicated, 1 mM glutamine, glutamate, arginine, or lysine was added to the EG medium for acidic adaptation and challenge.

RNA Isolation, cDNA Synthesis, and Quantitative Reverse Transcription (qRT)-PCR

RNA isolation and cDNA synthesis were carried out as described previously [2]. Gene expression was determined by qRT-PCR using Power SYBR Green PCR Master Mix (Applied Biosystems, CA) according to the manufacturer’s protocol, which was performed as described previously [2]. The primer sets used are listed in Table 1. The mRNA levels relative to those of the wild type were calculated using the comparative critical threshold method (2−∆∆CT), with rplU as an internal control gene as described previously [22].

Construction of the Chromosomal Translational Fusion and Other DNA Manipulations

The cadB::lacZ gene fusion was constructed on plasmids and then transferred in single copy to the chromosome of strain MC4100 by in vivo recombination as described previously [32]. A 975-bp cadB DNA fragment (from −796 to +179 relative to the translational start site) containing 59 codons was amplified by PCR using chromosomal DNA from strain MC4100 as template, primers cadL796 and cadR179, and Taq DNA polymerase (New England Biolabs, MA), using a Tm of 58 °C. The amplified DNA was digested with EcoRI and BamHI, and cloned into the vector pRS552 [38], yielding the translational fusion plasmids pCadB975, insert sequence of which was verified by sequencing. Strain MC4100 was subsequently transformed with the plasmid pCadB975; a transformant of strain MC4100 was infected with λ phage ARZ5 [32]; and the resulting lysate was used to transduce strain MC4100. Ampicillin-resistant, Lac+ lysogens were isolated by plating onto LB plates containing ampicillin and X-gal (40 μg/ml). Phages were prepared from these lysogens and used to reinfect strain MC4100 at a multiplicity of infection of 0.01. The resulting strain BHCB2 was then screened using restriction enzyme analysis to confirm the presence of the cadB: :lacZ fusion. The lysogens (at least six isolates) were then screened for the presence of a single copy of the recombinant ARZ5 phage by measuring the activity of phage-encoded β-galactosidase and β-lactamase. The lysogens displaying the lowest enzyme levels were assumed to carry a single prophage.

Plasmid CaCl2 transformation and phage P1 transduction to transfer mutations were performed according to standard procedures [35]

Western Blot Analysis

E. coli was cultured overnight in LB medium at 37 °C and then diluted 1:1,000 in EG medium with 5 mM lysine at pH 5.5 for acidic adaptation. After being anaerobically grown to the absorbance at 600 nm of 0.3, inside–out membrane vesicles were prepared, and Western blot analyses were carried out as described previously [40]. Rabbit polyclonal antibodies against CadB were kindly supplied by K. Igarashi. For Western blot analysis of CadB, membrane vesicles (20 μg protein) were separated by SDS–polyacrylamide gel electrophoresis on a 12 % acrylamide gel and transferred to a polyvinylidene fluoride membrane (Immobilon P, Millipore Bedford, MA). Antibodies were bound to proteins on the membrane. Staining was carried out with anti-rabbit IgG antibody conjugated with alkaline phosphatase. CadB protein was detected with ProtoBlot Western blot AP System (Promega).

β-Galactosidase Assay

β-galactosidase assays were performed as described previously [27]. Cells were cultured to log phase. The data were recorded in triplicate with more than three independent experiments.

Prediction of RNA Secondary Structure

The predicted secondary structure for full-length 5′ UTR of cadB was analyzed for minimal free energy and RNA secondary structure using the online MFOLD program prepared by M. Zuker [48].

Results

The Effects of ihfA and ihfB Deletions on the Induction of AR

Escherichia coli cells grown to stationary phase have been most used in the studies of AR induction, and such cells are more resistant to various stresses. To exclude the effect of IHF on stationary phase gene expression, cells growing exponentially were used in this study. The ihf-mutant strains were first cultured in EG medium at pH7.5 until the OD600 reached 0.3. Then, the cells were harvested by centrifugation, washed with EG medium at pH5.5, and suspended in the same medium at pH5.5. These cultures were adapted to the weakly acidic pH for 4 h to increase AR, and then survival at pH2.5 was measured.

Based on these data, only a small effect on the AR was observed in strains carrying ihfA or ihfB single deletion. However, survival of the strain lacking both IHFα and IHFβ subunits was 17-fold lower than that of the parent strain (Fig. 1a), indicating that IHF is required for the induction of AR.

Fig. 1
figure 1

The effect of ihf deletions on AR induction. The AR values for the wild type strainW3110 and its isogenic derivatives, BH32 (ihfA), BH33 (ihfB) and BH36 (ihfAihfB) were measured in EG medium after a 2 h acid challenge as described in Materials and Methods. No amino acids (a), Glutamine (gln) (b), glutamate (glu) (c), arginine (arg) (d) or lysine (lys) (e) was added at 1 mM as indicated in the figure. The means and standard deviations of duplicate determinations from at least three independent cultures are represented

Full size image

The Role of IHF in the Amino Acid-Dependent Induction of AR

It is well established that glutamate, glutamine, arginine, and lysine increase AR [11, 21]. We next examined if IHF was required for the induction of amino acid-dependent AR (AR2, AR3 and AR4). The cells were adapted at pH 5.5, and then challenged at pH 2.5 in the presence of 1 mM of each amino acid followed by AR measurement. The glutamate- and glutamine-dependent AR inductions showed no significant effects upon deletion of the two ihf genes (Fig. 1b, c). However, the arginine-dependent and lysine-dependent survivals at pH2.5 were diminished by 18 to 41-fold by the deletion of single or double ihf genes (Fig. 1d, e), indicating that IHF has a role in the induction of arginine- and lysine-dependent AR.

The Effect of IHF Deletion on the Transcriptions of AR Structural Genes

The genes encoding amino acid decarboxylases and conjugate antiporters are known to play essential roles in the amino acid-dependent AR systems [21]. To test if IHF affects their expression, the mRNA levels of several genes were measured quantitatively by real-time PCR, and the relative levels of the ∆ihfA∆ifhB strain to those of the wild type strain were calculated as described in Materials and methods. No significant change caused by the IHF mutation was observed with adiC, whereas the adiA transcript level was nine-fold lower in the ∆ihfA∆ifhB strain than in the wild-type strain (Fig. 2), in agreement with previous lacZ transcriptional fusion data [33]. Thus, it seems that IHF regulates arginine-dependent AR3 induction by activating the transcription of the gene adiA, but not adiC. The levels of the gadA, gadBC, and ybaS (glutaminase A gene) transcripts were approximately two-fold higher in the ∆ihfA∆ifhB strain than those in the wild-type strain (Fig. 2). The ybaS gene was shown to be required for glutamine-dependent AR induction [2]. The slightly increased mRNA levels of gadABC and ybaS upon IHF deletion are insufficient to affect the induction of the glutamate- or glutamine-dependent AR systems.

Fig. 2
figure 2

The effect of IHF mutation on the transcription of AR induction genes. Strain W3110 (wild type) and BH36 (ihfA ihfB) were grown in EG medium at pH 5.5 for 4 h and harvested before acidic challenge (pH2.5). The EG medium was supplemented with 1 mM of glutamine, glutamate, arginine, or lysine to assess the transcriptional levels of ybaS; gadA, B, and C; adiA and adiC; or cadBA, respectively. Real-time PCR was performed, and the levels of Strain BH36 expression (black bars) relative to those of the wild type strain (open bars) are represented, as described in the Materials and Methods. The data are the averages ± standard deviations of three independent experiments

Full size image

More interestingly, transcription of cadBA also appeared to be somewhat increased by loss of IHF (Fig. 2). However, the relatively high level of cadBA mRNA in the ∆ihfA∆ifhB strain cannot account for low-survival ability of the mutant strain after acidic challenge in the presence of lysine.

IHF Exerts Translational Control of CadB Expression

To determine if IHF induces lysine-dependent AR by control of the CadBA expression at acidic pH, we performed Western blot analysis to detect CadB expression. Under lysine induction conditions, the CadB protein levels were significantly lower in both the ihf single- and double-mutant strains compared with the wild-type strain (Fig. 3a). Lysine addition induced the CadB expression in the WT cell, consistent with a previous report [30]. To confirm the reduced CadB expression in the IHF- mutant cells, we constructed a translational fusion with lacZ as a reporter gene and measured the β-galactosidase activity to examine cadB translation. The level of β-galactosidase activity in the ∆ihfA∆ifhB strain was eight-fold lower than that in the wild-type strain at acidic pH with the addition of lysine. Similar fold decreases were observed in the ihfA- or ihfB-single- mutant cells (Fig. 3b), suggesting that IHF is essential for the translational control of cadB expression.

Fig. 3
figure 3

The effect of IHF mutation on cadB translational expression. a Western blot analysis of CadB expression. BHCB2 (wild type), BHCB34 (ihfA), BHCB35 (ihfB) and BHCB36 (ihfA ihfB) were anaerobically grown with 5 mM lysine at pH 5.5 and western blot analysis was performed using 20 μg of protein of the inside-out membrane vesicles as described in the Materials and Methods. b The effect of ihf deletions on β-galactosidase expression from a chromosomal cadB::lacZ translational fusion. Cells were grown anaerobically as in Fig. 3a and β-galactosidase activity was measured. The data are the averages ± standard deviations of three independent experiments

Full size image

Discussion

AR can help E. coli cells to overcome the challenge posed by extreme acidic environments. When they inevitably have to pass through the gastric acid in the stomach (pH 2.5), which serves as a natural antibiotic barrier, E. coli can survive in this hostile acidic condition. Therefore, understanding the complex regulatory mechanisms and pathways of AR response is crucial to developing strategies for clinical E. coli infection. In this study, we showed that IHF induces arginine- and lysine-dependent AR. To our knowledge, this is the first report showing the function of IHF in the control of AR induction.

It is generally considered that IHF functions as a transcriptional dual regulator and is involved in a variety of processes including transcription of certain promoters. However, we here reported that IHF also acts as a translational regulator. Our current data suggest that IHF induces the lysine-dependent AR by activating the cadBA translation at acidic pH. But it is not clear that the regulation by IHF is direct or indirect. Two possible mechanisms as to how IHF stimulates cadB translation can be proposed: either IHF modifies the RNA secondary structure to facilitate ribosome binding, or it modulates binding of other regulatory proteins as a chaperoning role, or both. Actually, the proposed secondary structure of 5′ UTR of the cadB mRNA showed that a long mRNA stem–loop structure forms at the ribosome-binding site (Fig. 4). IHF could bind the cadB 5′ UTR to prevent the formation of RNA duplex structures that occlude the ribosome binding site of the cadB transcript. This question is currently being investigated. Thus, IHF might also be a RNA-binding protein that stimulates protein expression and affects cellular processes.

Fig. 4
figure 4

Prediction of RNA secondary structure of the cadB 5′ UTR. The structure and ∆G 0 for 5′ UTR were displayed using the online MFOLD program [48], with the default settings and folding temperature at 37 °C. Designations: +1, transcriptional start site; RBS, ribosome binding site; Met, initiator methionine. The transcriptional start site was mapped by Watson N [44]

Full size image

In this study, the IHF deletion decreased survival in medium without the addition of amino acids (Fig. 1a), suggesting that IHF regulates expression of other genes required for survival at an extremely low pH. We also reported previously that HU has a similar role in AR induction [2]. Actually, IHF and HU monomers share ~30 % sequence identity and are proposed to be related in both structure and function [1]. Furthermore, the HU/IHF family of proteins consists of orthologs that share significant sequence identity. Thus, our results support the importance of nucleoid-associated proteins function during environmental adaptation. Compounds that inhibit HU/IHF activity could be effective antimicrobials that are against multiple species of bacteria.