INTRODUCTION

Top Abstract Introduction Materials and methods Results and discussion References

In microorganisms growing on acetate as the sole carbon source, isocitrate can be catabolized either by the Krebs cycle or by the glyoxylate bypass (18). In Escherichia coli, this branch point is regulated by the reversible phosphorylation and the concomitant inactivation of the NADP⁺-dependent isocitrate dehydrogenase (IDH; EC 1.1.1.42) (9, 19, 26). Such a phosphorylation or dephosphorylation process plays a pivotal role in cell survival by connecting, via the glyoxylate pathway, the metabolism of two-carbon compounds with gluconeogenesis and by allowing bacteria to draw energy through the oxidative decarboxylation steps of the Krebs cycle.

In addition to posttranslational control, IDH is also regulated at the level of expression of its gene (icd). It has been demonstrated that under anaerobic cell growth conditions, the icd gene is negatively controlled by both the ArcAB system and the Fnr factor (15). Recently, Chao et al. (5) provided direct evidence of the negative role played by ArcA in icd gene regulation by demonstrating in vitro binding of this repressor at a target operator site overlapping the icd promoter. Also, previous work has shown that the Cra protein (initially called FruR) exerts positive control over a number of genes and operons encoding biosynthetic and oxidative enzymes, including icd (reviewed by Saier and Chin [30]). In the latter case, a single Cra-binding site was detected by electrophoretic mobility shift analysis (EMSA) in the DNA fragment encompassing the regulatory region of the icd gene (28).

This paper focuses on the role of protein Cra in transcription of the icd gene. Using an in vitro transcription approach, we found that in addition to its main promoter (P1), the icd control region contains a second promoter whose activation is dependent on the action of Cra (P2). The start points relative to these promoter sites were mapped by primer extension analysis, and then the precise contacts between RNA polymerase (RNAP) and its promoters and between Cra and its DNA operator were analyzed by the base removal method and the DNase I footprinting technique, respectively. Finally, testing of specific point mutations in the  subunit of RNAP revealed that a number of RNAP-DNA contacts play a key role in transcription of the icd gene.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and methods Results and discussion References

Proteins. Recombinant active Cra protein with a His₆ tag at its C-terminal end was prepared from E. coli BL21(DE3) cells (33) harboring the overproducing plasmid pJCD2 (6). Wild-type and mutant  subunits of RNAP were reconstituted in vitro from separately purified subunits as described previously (14). The specific activity of wild-type and mutant -subunit holoenzymes was determined by measuring the level of poly(dA-dT)-dependent poly(AU) synthesis.

Plasmids. (i) Plasmid pJFC2. The icd promoter region (EMBL accession no. J02799) (34) was generated by PCR amplification from E. coli K-12 chromosomal DNA. The synthetic oligonucleotide primers used in this reaction were P_icd-EcoRI (5'-TATGAATTCAGGTTTACGCC-3') and P_icd-PstI (5'-TATCTGCAGGGTGATCTT-3'). Following PCR, the fragment was restricted with EcoRI-PstI and cloned in fusion with the lacZ gene into the compatible sites of the vector pNM481 (21) to create plasmid pJFC2. Plasmid pJFC2 was also used to generate radioactively labeled DNA fragments for gel retardation, base removal interference, and DNase I footprinting studies. In all cases, after restriction by EcoRI-PstI, the DNA fragment carrying the icd promoter fragment was end labeled at the EcoRI site by [-³²P]dATP (3,000 Ci/mmol) (bottom strand), using the Klenow fragment of DNA polymerase.

(ii) Plasmid pJFC1. pJFC1, derived from the pUC19 vector (GenBank accession no. X02514), was used as the template for in vitro transcription studies. It contained the EcoRI-PstI icd promoter-bearing DNA fragment inserted into the corresponding sites of pUC19, upstream of the transcriptional termination signal T1T2 of the rrnB operon (1).

Primer extension analysis. RNA was isolated from E. coli [pJFC2] cells essentially as described by Reddy et al. (29). The oligonucleotide primer 5'-TGCATATGCGTTTGCGTCCTGCGATACGGA-3' (250 pmol) was end labeled according to the standard procedure (31), using 30 µCi of [-³²P]ATP (3,000 Ci/mmol) and T4 polynucleotide kinase (Promega Corp.). Primer extension reaction was performed as described elsewhere (25), using 45 µg of total cellular RNA. Extension products were resolved by electrophoresis in a 6% (wt/vol) polyacrylamide-7 M urea gel and visualized by autoradiography.

In vitro transcription assay. Single-round in vitro transcription experiments were performed with template plasmid pJFC1 as follows. Five picomoles of supercoiled plasmid pJFC1 was preincubated for 25 min at 30°C with 1 pmol of either wild-type or mutant -subunit RNAPs in a 20-µl assay mixture containing 50 mM Tris-acetate (pH 8.0), 100 mM potassium acetate, 8% (vol/vol) glycerol, 0.1 mM EDTA, 8 mM magnesium acetate, 0.1 mM dithiothreitol, and 500 U of RNAsin per ml. When required, 25 pmol of active Cra protein was added. Transcription reactions were initiated by the addition of 0.2 mM each ATP, GTP, and CTP, 0.01 mM UTP, 2 µCi of [-³²P]UTP (800 Ci/mmol), and 100 µg of heparin per ml. After 10 min of incubation at 37°C, the reactions were blocked with 1 volume of gel loading buffer, the mixtures were heated at 65°C for 3 min and analyzed in a 6% polyacrylamide-7 M urea gels. Quantification of bands on the gel was achieved by densitometric analysis of the autoradiograms.

EMSA. Gel shift assays were performed as previously described (10). Typically, a 5'-end-labeled DNA fragment (10⁵ cpm) was incubated with either RNAP (100 nM) or active Cra protein (50 nM) for 10 min at 25°C in 20 µl of DNA binding buffer [12 mM HEPES-NaOH (pH 7.9), 4 mM Tris-HCl (pH 7.9), 95 mM KCl, 1 mM EDTA, 1 mM dithiothreitol, 9% [vol/vol] glycerol, 0.02% [vol/vol] Nonidet P-40, 2 µg of poly(dI-dC) · poly(dI-dC), 10 µg of bovine serum albumin per ml]. Complexes were resolved from free DNA on a 4% nondenaturing polyacrylamide gel and electrophoresed in Tris-glycine buffer (50 mM Tris, 384 mM glycine, 2.1 mM EDTA) at 200 V for 1 to 2 h at 25°C. Radioactive bands were visualized by autoradiography.

DNase I footprinting. DNase I footprinting was performed in 100 µl of DNA binding buffer (see above) with purified Cra protein (10 or 50 nM) and the 5'-end-labeled DNA fragment (10⁵ cpm) that contained the icd promoter. Following incubation of the reaction mixture for 10 min at 25°C, 1.6 U of DNase I (Stratagene) in the presence of 2 mM CaCl₂ and 6 mM MgCl₂ was added; digestion was allowed to proceed for 50 s and was stopped by the addition of 10 mM EDTA. After precipitation with ethanol, nucleic acids were dissolved in formamide dye mix and heated at 80°C for 3 min prior to electrophoresis in a 6% acrylamide-7 M urea sequencing gel.

Base removal experiments. Base removal experiments were performed by the procedure described by Brunelle and Schleif (2) as modified by Nègre et al. (23).

	RESULTS AND DISCUSSION

Top Abstract Introduction Materials and methods Results and discussion References

In vitro transcription of the icd gene. An in vitro transcription assay was carried out with the supercoiled plasmid pJFC1, which contains the icd regulatory region (Fig. 1A) cloned upstream of the strong rrnB T1 and T2 terminators (24). The products obtained from single-round transcription experiments were analyzed by electrophoresis in a 6% sequencing gel and detected by autoradiography.

View larger version (38K): [in this window] [in a new window]   FIG. 1.   (A) Nucleotide sequence of the icd regulatory region (from EMBL accession no. J02799). The 10 and 35 regions of the icd P1 and icd P2 promoters are underlined; locations of the transcription initiation sites of the two promoters are indicated by arrows. The high-affinity Cra-binding site (Cra box) is centered at position 76.5 relative to the P2 transcription initiation site. (B) In vitro transcription from icd promoters with wild-type RNAP. Transcription assays were performed with plasmid pJFC1 as the template. The positions of the RNA-I (nt 107 to 108), icd P1 (nt 225), and icd P2 (nt 272) transcripts are shown by arrows.

Location of the icd mRNA start sites. Primer extension analysis was performed to identify the transcriptional start points of in vivo transcription from icd P1 and icd P2. Total cellular RNA was primed with a radiolabeled oligonucleotide complementary to the icd regulatory sequence, and this primer was extended by using reverse transcriptase. The primer used for cDNA synthesis was also used in a DNA sequencing reaction to determine the length of each extension DNA product.

View larger version (24K): [in this window] [in a new window]   FIG. 2.   Mapping of the icd transcriptional start sites by primer extension analysis. A 32P-end-labeled oligonucleotide primer (106 cpm) was annealed to 45 µg of total RNA isolated from E. coli JM109[pJCD2] cells and extended by using deoxyribonucleoside triphosphates and avian myeloblastosis virus reverse transcriptase. The cDNA products were separated by electrophoresis on a 6% polyacrylamide-7 M urea gel. The mobility of cDNA products was compared to a DNA sequencing ladder (GATC) generated with the same primer as that used for the primer extension reactions. Arrows indicate the nucleotides corresponding to the lengths of the extension products.

Interaction of RNAP with the icd promoter region. In bacteria, the transcription process initiates through the formation of a closed complex between RNAP and the promoter site, followed by the isomerization to a transcription-competent open structure in which the DNA encompassing residues from approximately 12 to +2 (with respect to the transcription start site at +1) represents the melting domain. To assess the role of individual base residues in the specific interaction between RNAP and the promoter, the base removal interference technique (2) has been found to be a powerful approach (17, 23, 36).

View larger version (31K): [in this window] [in a new window]   FIG. 3.   Identification of RNAP-icd DNA contacts by base removal footprinting. The 328-bp icd regulatory region, 32P-labeled at the EcoRI site (Fig. 1A), was treated with either formic acid to remove G+A or hydrazine to remove C+T and then incubated with RNAP (approximately 10 nM active protein) prior electrophoresis on a 4% preparative polyacrylamide gel at high ionic strength. Free and complexed DNA molecules were then subjected to piperidine cleavage and analyzed on a 6% sequencing gel to reveal the positions of the interfering modifications. Lanes: B, modified DNA isolated from complexes; F, DNA that had dissociated or was free of complex.

Cra binding to the icd promoter region. The binding affinity of the active Cra factor for the icd promoter region was measured by the EMSA (10). The radiolabeled 328-bp DNA fragment carrying both icd P1 and icd P2 was incubated with increasing concentrations of Cra, and the corresponding complexes were resolved from free DNA by gel electrophoresis and quantified after autoradiography. At the concentrations of Cra used in these experiments, single stable complexes could be resolved by EMSA (Fig. 4A), which indicated that Cra could interact in a specific manner with one operator site in the icd regulatory region. The bandshift assay was performed at a low concentration of DNA (4 pM) such that the approximate free protein concentration equaled the total protein concentration; thus, at half-maximal saturation, there was >10-fold molar excess of protein relative to nucleic acid (4). The equilibrium binding constant, K_d, could therefore be derived as the total protein concentration at half-maximal saturation. The K_d value for Cra binding to the icd regulatory region at pH 7.9 and 25°C was about 10 nM, which indicated a relatively high binding affinity of Cra for this particular region of DNA.

View larger version (30K): [in this window] [in a new window]   FIG. 4.   Characterization of the Cra-icd regulatory region complex. (A) Relative strength of binding of Cra to the icd regulatory region. The 328-bp 32P-labeled icd DNA (4 pM) was incubated at 25°C with increasing (from 0 to 250 nM) concentrations of Cra protein and analyzed by EMSA. (B) Cra-mediated protection of the icd regulatory region against digestion by DNase I. Target DNA was incubated in the absence (lane 1) or in the presence of 10 and 50 nM Cra protein (lanes 2 and 3, respectively). Samples were separated on a 6% polyacrylamide sequencing gel. G+A shows the Maxam-Gilbert sequencing ladder of the probe; arrows delineate the footprint produced by Cra.

Effects of Ala substitutions in the CTD of RNAP on icd transcriptional activation. To determine the importance of the CTD in icd P2 transcription, we treated a series of alanine mutant RNAPs containing point mutations in this region (22) for transcription in the presence of plasmid pJFC1. The results (Fig. 5) showed that each mutant RNAP tested could initiate transcription from the icd P1 promoter at nearly the same rate as the wild-type enzyme. On the other hand, the transcription initiated from the icd P2 promoter, with or without Cra, was considerably reduced when incubation was carried out with mutant RNAPs containing substitution L262A, R265A, and N268A ([L262A], [R265A], and [N268A]), whereas a less drastic effect was observed with mutant RNAPs [L260A], [C269A], and [K297A]. Moreover, comparison of the icd P2/icd P1 transcription ratio, in both the presence and absence of Cra, showed that both basal transcription (Fig. 5A) and the Cra-dependent activation of transcription (Fig. 5B) were affected concomitantly by these mutations. This result strongly supports the hypothesis that during formation of the closed complex, basal transcription initiation as well as Cra-dependent activation of transcription strictly rely on contacts established between the  subunit and the proximal DNA region upstream from the promoter.

View larger version (51K): [in this window] [in a new window]   FIG. 5.   In vitro transcription by wild-type (WT) and mutant RNAPs. Reconstituted RNAP holoenzymes containing Ala substitutions in the CTD were assayed for transcription in the presence (A) or absence (B) of Cra. Single-round transcriptions were performed by using supercoiled plasmid pJFC1 as the template in the presence of mutant RNAP and protein Cra in a molar ration of 1:1. Transcripts were separated in a 6% sequencing gel and detected by autoradiography. The ratio of the transcriptional activity at icd P2 relative to that at icd P1 is indicated below each gel.

View larger version (26K): [in this window] [in a new window]   FIG. 6.   Characterization of the complexes between wild-type (WT) or mutant RNAPs and the icd P2 promoter. A 120-bp 32P-labeled DNA carrying the icd P2 promoter region was incubated at 25°C with a decreasing concentration of RNAP and analyzed by EMSA. Positions of the protein-DNA complex and free DNA are indicated.