Effects of integration host factor and DNA supercoiling on transcription from the ilvPG promoter of Escherichia coli.

Integration host factor (IHF) activates transcription from the ilvPG promoter by severely distorting the DNA helix in an upstream region of a supercoiled DNA template in a way that alters the structure of the DNA in the downstream promoter region and facilitates open complex formation. In this report, the in vivo and in vitro influence of DNA supercoiling on transcription from this promoter is examined. In the absence of IHF, promoter activity increases with increased DNA supercoiling. In the presence of IHF, the same increases in superhelical DNA densities result in larger increases in promoter activity until a maximal activation of 5-fold is obtained. However, the relative transcriptional activities of the promoter in the presence and absence of IHF at any given DNA superhelical density remains the same. Thus, IHF and increased DNA supercoiling activate transcription by different mechanisms. Also, IHF binds with equal affinities to its target site on linear and supercoiled DNA templates. Therefore, IHF binding does not activate transcription simply by increasing the local negative supercoiling of the DNA helix in the downstream promoter region or by differential binding to relaxed and supercoiled DNA templates.

Integration host factor (IHF) activates transcription from the ilvP G promoter by severely distorting the DNA helix in an upstream region of a supercoiled DNA template in a way that alters the structure of the DNA in the downstream promoter region and facilitates open complex formation. In this report, the in vivo and in vitro influence of DNA supercoiling on transcription from this promoter is examined. In the absence of IHF, promoter activity increases with increased DNA supercoiling. In the presence of IHF, the same increases in superhelical DNA densities result in larger increases in promoter activity until a maximal activation of 5-fold is obtained. However, the relative transcriptional activities of the promoter in the presence and absence of IHF at any given DNA superhelical density remains the same. Thus, IHF and increased DNA supercoiling activate transcription by different mechanisms. Also, IHF binds with equal affinities to its target site on linear and supercoiled DNA templates. Therefore, IHF binding does not activate transcription simply by increasing the local negative supercoiling of the DNA helix in the downstream promoter region or by differential binding to relaxed and supercoiled DNA templates.
The chromosomal DNA of the bacterium Escherichia coli is highly compacted and negatively supercoiled (1)(2)(3). The supercoiled state of the E. coli chromosome is primarily maintained by the opposing activities of two topoisomerases, DNA gyrase and DNA topoisomerase I. DNA gyrase, composed of two subunits encoded by the gyrA and gyrB genes, introduces negative supercoils into DNA in an ATP-dependent manner; whereas, DNA topoisomerase I, the product of the topA gene, removes negative supercoils from DNA by an ATP-independent mechanism (1). The supercoiled state of the chromosome is known to effect the activity of many promoters. The activity of some promoters is greatly diminished when DNA is relaxed, while the activity of others is unaffected or even enhanced (1).
The ilvGMEDA operon of E. coli is required for the biosynthesis of the branch chained amino acids, L-isoleucine and L-valine (4). This operon is preceded by a strong 70 -promoter, ilvP G . Transcription from this promoter is activated by two upstream activating sequences, UAS1 and UAS2 (Fig. 1).
IHF forms a higher-order protein-DNA complex in UAS1 that facilitates unwinding of the DNA helix in the Ϫ10 hexanucleotide region of the downstream ilvP G promoter, and this binding is accompanied by an increase in the rate of opencomplex formation (7). These and other observations (see "Discussion") coupled with the observations that IHF-mediated activation occurs in the absence of specific protein interactions between IHF and RNA polymerase, and that IHF-mediated activation requires a negatively supercoiled DNA template, support the hypothesis that IHF activates transcription from this promoter by an allosteric DNA mechanism that is influenced by the superhelical state of the DNA template (7). In this report, we examine the effects of DNA supercoiling on IHFmediated activation. We show that the activity of the ilvP G promoter increases with increased negative supercoiling and that this sensitivity to superhelical density is enhanced in the presence of IHF. We further show that IHF and DNA supercoiling influence transcription from this promoter by different mechanisms, and that activation is not the consequence of differential binding of IHF to relaxed and supercoiled DNA templates.

MATERIALS AND METHODS
Chemicals and Reagents-Restriction endonucleases, T4 DNA ligase, and T4 polynucleotide kinase were purchased from New England Biolabs. E. coli RNA polymerase, pancreatic RNasin, and DNase I were purchased from Boehringer Mannheim. Radiolabeled nucleotides were obtained from DuPont NEN. DNA probes were radiolabeled using a nick translation kit purchased from Amersham Corp. DNA sequencing was performed using the Sequenase kit of U. S. Biochemicals. DNA oligonucleotides were synthesized on an Applied Biosystems PCR Mate DNA synthesizer. Integration host factor was purified in this laboratory by the method of Nash et al. (13).
Plasmids and Bacterial Strains-Plasmid DNA isolation and all recombinant DNA manipulations were carried out using standard methods (14). Plasmids and bacterial strains used in this study are described in Table I.
␤-Galactosidase Assays-Cells were grown at 37°C in logarithmic phase to a culture density of 0.5 to 0.7 OD 600 units in M63 minimal salts media containing 0.4% glucose (15). Cell growth was arrested by chilling the culture on ice. ␤-Galactosidase activities were assayed by measuring o-nitrophenyl ␤-D-galactoside hydrolysis in SDS-chloroform permealized cells. ␤-Galactosidase activities were measured at four different time points and two extract concentrations under conditions where the assay was linear with respect to time and extract concentration. Rates of o-nitrophenol formation were determined by a linear regression analysis of an o-nitrophenol versus time plot and specific activities were calculated according to the method of Miller (15).
Determination of IHF Binding Affinity on Linear and Supercoiled DNA Templates-The binding affinity of IHF to its target site in the ilvP G promoter regulatory region was determined on a linear DNA template by gel mobility shift assays. A 471-bp EcoRI-HindIII DNA fragment containing the ilvP G promoter region from ilv bp position Ϫ360 to ϩ60 (16) was isolated from plasmid pABC209 (Table I) and radiolabeled at each 5Ј end with T4 polynucleotide kinase and 10 Ci of [␥-32 P]ATP (3,000 Ci/mmol). The radiolabeled DNA (1 ϫ 10 Ϫ11 M final concentration) was preincubated with purified IHF in a 20-l assay mixture (40 mM Tris-HCl (pH 8.0), 4 mM MgCl 2 , 70 mM KCl, 0.1 mM EDTA, 0.1 mM dithiothreitol, and 25 g/ml herring sperm DNA). The free IHF concentration in each sample was assumed to be the same as the total IHF concentration since the DNA template concentration was significantly lower than that of the total protein concentration. The DNA fragments and IHF were incubated at 25°C for 20 min and the free and IHF-bound DNA fragments were separated by electrophoresis on a 5% polyacrylamide gel (4.83% acrylamide, 0.17% N,NЈ-methylenebisacrylamide) in TAE buffer (40 mM Tris acetate (pH 8.0), 1 mM EDTA (14)). Electrophoresis was performed at 25°C with constant current (20 mA) for 3 h. Free and IHF-bound DNA fragments were visualized by autoradiography following the exposure of the dried gels to Kodak XAR-5 film at Ϫ70°C in the presence of a Cronex Quanta III intensifying screen (DuPont). Quantitation of band intensity on autoradiographic film was performed utilizing the public domain NIH IM-AGE gel quantitation software (ftp://zippy.nimh.nih.gov). Determination of equilibrium dissociation constants (K D ) was performed as described by Brenowitz et al. (17). According to this method the binding curve is described by the Langmuir isotherm, Y ϭ k[P]/1ϩk [P]. The equilibrium binding data were analyzed by a non-linear least squares parameter estimation method. The algorithm for this analysis (18) uses a variation of the Gauss-Newton procedure (19) to determine the best fit, model-dependent, parameter values corresponding to a minimum in the variance of each data point. The confidence levels for the curve fits reported correspond to approximately 1 standard deviation (65% confidence). In fitting the data to the equations, the substitution, ⌬G ϭ -RTln K, was made so that the ⌬G values for each experiment were the actual curve fit parameters.
Quantitative DNase I footprinting assays were performed to determine the binding affinity of IHF to its target site on a negatively supercoiled DNA template. Supercoiled plasmid DNA (1 ϫ 10 Ϫ11 M final concentration) was incubated for 20 min at 25°C with purified IHF in the same assay mixture used for the gel mobility shift assays. The IHF-DNA mixture was treated with 5 ng of DNase I for exactly 2 min to insure single-hit kinetic conditions (17). DNase I reactions were stopped by placing the samples in boiling water for 5 min. The sites of protection from DNase I cleavage were mapped by primer extension using Taq DNA polymerase and an ilv-specific oligonucleotide, OL-132 (40 pmol), that anneals to ilv bp positions Ϫ155 to Ϫ132 (16). Thirty cycles of primer extension were performed in a Perkin Elmer DNA Thermal Cycler Model 480 (1 min at 95°C (denaturing), 1 min at 55°C (reannealing), and 1 min at 72°C (extension)). The amplified primer extension products were separated by electrophoresis on an 8% denaturing polyacrylamide gel (7.6% acrylamide, 0.4% N,NЈ-methylenebisacrylamide) containing 8 M urea in TBE buffer (90 mM Tris borate (pH 8.0), 1 mM EDTA (14)) and visualized by autoradiography following exposure of the gels to Kodak XAR-5 film at Ϫ70°C in the presence of a Cronex Quanta III intensifying screen (DuPont). For quantitation, the band intensities in the protected region of each lane were normalized to the intensity of the band at bp Ϫ69. The equilibrium dissociation constants were determined by the methods of Brenowitz et al. (17) as described above.
In Vitro Transcriptions-In vitro transcription reactions were performed according to the procedures of Hauser et al. (20), with closedcircular supercoiled plasmid pDH⌬wt (Table I) Reactions were terminated after 3 and 6 min by removing a 15-l sample and adding it to 15 l of stop solution (95% formamide, 0.025% bromphenol blue, 0.025% xylene cyanol). The reaction products were separated by electrophoresis on an 8% denaturing polyacrylamide gel (7.6% acrylamide, 0.4% N,NЈ-methylenebisacrylamide) containing 8 M urea in TBE buffer (14) and visualized by autoradiography following exposure of the gels to Kodak XAR-5 film at Ϫ70°C in the presence of a Cronex Quanta III intensifying screen (DuPont).
Generation of Plasmid DNA Topoisomers-10 g of plasmid pDH⌬wt (Table I) was treated with 20 units of Drosophila melanogaster topoisomerase II in a 60-l reaction mixture (10 mM Tris-HCl (pH 7.9), 50 mM NaCl, 50 mM KCl, 5 mM MgCl 2 , 0.1 mM EDTA, 15 g/ml bovine serum albumin, 1 mM ATP and ethidium bromide (0 -20 M)) for 4 h. Each plasmid DNA sample was extracted three times with phenol to remove the ethidium bromide, precipitated with 2 volumes of isopropyl alcohol, and resuspended in 20 l of water. The plasmid DNA topoisomers were resolved by electrophoresis on four, 1.4%, agarose gels in TAE buffer (14) containing 0.006, 0.02, 0.04, and 0.08 g/ml of ethidium bromide, respectively. The average linking number difference of the DNA plasmid in each sample (⌬Lk) was determined by the band counting methods of Keller (21) and Singleton and Wells (22). The average superhelical density () was calculated using the equation ϭ Ϫ10.5⌬Lk/N, where N is the number of bp in the plasmid (pDH⌬wt contains 4203 bp).

RESULTS
In Vitro Effect of DNA Supercoiling on Transcription from the ilvP G Promoter-A set of DNA topoisomers of the plasmid pDH⌬wt (Table I), which ranged in average linking number deficiencies (⌬Lk) from 0 to Ϫ48, corresponding to negative superhelical densities of ϭ Ϫ0.00 to Ϫ0.14, were prepared as described under "Materials and Methods" (Fig. 2). The transcriptional activity of the ilvP G promoter on DNA templates with different superhelical densities was determined by measuring the rate of ilv specific transcript production in an in vitro transcription reaction (Fig. 3). The in vitro transcription reactions were performed with a minimally saturating concentration of RNA polymerase under conditions where the rate of transcript formation was directly proportional to the DNA template concentration (data not shown). Under these experimental conditions, the transcriptional activity of the ilvP G promoter is below detectable levels on a completely relaxed DNA template ( ϭ Ϫ0.00), is detectable at a superhelical DNA density of about ϭ Ϫ0.01, increases in the physiological range of superhelical DNA densities from ϭ Ϫ0.03 to Ϫ0.09, and reaches a maximal transcription rate at a superhelical DNA density of about ϭ Ϫ0.09 (Figs. 3 and 4). These results demonstrate that the transcriptional activity of the ilvP G promoter is intrinsically sensitive to the superhelical density of the DNA template.
In Vitro Effect of DNA Supercoiling and IHF on Transcription from the ilvP G Promoter-When a nearly saturating concentration of IHF (20 nM) was included in in vitro transcription FIG. 1. Nucleotide sequence of the ilvP G promoter-regulatory region of the ilvGMEDA operon of E. coli. The nucleotide sequence of the ilvP G promoter region from bp Ϫ185 to ϩ1 is shown. The nucleotides are numbered corresponding to the in vivo transcriptional start site from the ilvP G promoter. The IHF core binding site, 5Ј-AAACAA-CAATTTA-3Ј, in the upstream activating sequence UAS1 is located between bp Ϫ82 and Ϫ96 (16). UAS2, located between bp Ϫ41 and Ϫ59, contains a set of helically phased adenine residues. The proposed Ϫ10 and Ϫ35 hexamer nucleotide regions of the ilvP G promoter are underlined.
reactions, the effect of superhelical DNA density on the rate of transcription from the ilvP G promoter was similar to that observed in the absence of IHF. That is, the superhelical DNA density that produced half-maximal promoter activity in the presence of IHF was the same as the superhelical DNA density that produced half-maximal promoter activity in the absence of IHF (Figs. 4 and 5). However, in the presence of IHF, increases in superhelical DNA densities resulted in larger increases in promoter activity until a maximal activation of 5-fold was obtained at a superhelical DNA density of ϭ Ϫ0.09 (Figs. 4 and 6). At higher superhelical DNA densities, no further activation was observed and at superhelical DNA densities below ϭ Ϫ0.03 little, if any, IHF-mediated activation was apparent. These data suggest that IHF does not affect the overall response of the ilvP G promoter to changes in superhelical DNA density; rather, in the presence of IHF, promoter activity is more sensitive to small changes in the superhelical density of the DNA template in the physiological range of ϭ Ϫ0.03 to Ϫ0.09. Furthermore, the observation that in the absence of IHF, transcriptional activity did not continue to increase beyond a superhelical DNA density of ϭ Ϫ0.09 (Fig. 4) suggests that the activation role of IHF cannot be replaced simply by increasing the superhelical density of the DNA template. Thus, IHF and increased negative supercoiling activate transcription from the ilvP G promoter by different mechanisms.
In Vivo Effect of DNA Supercoiling on Transcription from the ilvP G Promoter in the Absence and Presence of IHF-To determine if changes in DNA supercoiling affect promoter activity in vivo in the same manner as in vitro, promoter activities were measured in gyrase-(gyrB) and topoisomerase (topA)-deficient strains containing a functional or a mutated himA (IHF) gene. The changes in in vivo superhelical DNA densities effected by these mutations were monitored by measuring the superhelical densities of a reporter plasmid (pDH⌬wt; Table I) harvested from appropriate mutant strains during mid-log growth as described under "Materials and Methods" (Table II). The tran- scriptional activity of the ilvP G promoter was determined in each strain by measuring the expression of a lacZ reporter gene transcriptionally fused to the ilvP G promoter and integrated into the bacterial chromosome in single copy. The results presented in Table II show that, like in vitro, the in vivo expression of the ilvP G promoter in the presence of IHF is proportional to the degree of negative DNA supercoiling in the physiological range of superhelical densities ( ϭ Ϫ0.03 to Ϫ0.09; Ref. 23). For example, the gyrB mutation decreases the superhelical density of a reporter plasmid 30% and decreases promoter activity a comparable 37% (compare strains IH-100 and IH-200). Also, the topA mutation increases the superhelical density of a reporter plasmid by 13% and the expression of the ilvP G promoter by 14% (compare strains IH-100 and IH-750). The data in Table II further show that the himA mutation causes a 3-fold drop in reporter gene expression (IHF-mediated activation) without significantly affecting in vivo DNA supercoiling. This result is consistent with the conclusion that IHF and DNA supercoiling affect ilvP G promoter activity by differ-ent mechanisms. However, in the double mutants (himA,gyrB and himA,topA) a proportional in vivo and in vitro correlation between ilvP G expression and superhelical DNA density was not observed. For example, in the himA,topA strain ilvP G expression is nearly twice as high as it is in a himA strain (compare strains IH-105 and IH-755); whereas, it is only slightly higher in a topA strain containing a functional IHF gene than it is in a wild type strain (compare strains IH-100 and IH750). This result suggests that, at high levels of chromosomal DNA supercoiling required for maximal IHF activation, the activating effects of other nonspecific in vivo chromosomal organizer proteins, such as HU or HN-S, might become apparent (24 -26). Also, in the gyrB,himA strain a much lower promoter activity than expected from the in vitro data was observed (compare strains IH-105 and IH-205). This might be the consequence of a much lower in vivo chromosomal superhelical DNA density in a gyrB,himA strain than in a gyrB strain. This supposition is supported by the reports of Friedman et al. (27,28) that although DNA is only slightly more relaxed in a himA strain than in a wild type strain, it is much more relaxed in a himA,gyrB strain than in a gyrB strain. IHF Binding Affinity on Topologically Relaxed and Negatively Supercoiled DNA Templates-Gel mobility shift assays were employed to measure the equilibrium dissociation constant K D of IHF to its target site in the ilvP G promoter region. A 32 P-end-labeled, 471-bp DNA fragment containing the ilvP G promoter region sequence from bp positions Ϫ360 to ϩ6 (16) was used for these assays (Fig. 7). Since the DNA concentration used in these assays (ϳ1 ϫ 10 Ϫ11 M) was significantly lower than the IHF concentrations (0.07 to 9.0 ϫ 10 Ϫ9 M), the free and total IHF concentrations were assumed to be the same. The IHF binding isotherm for its ilvP G target site, obtained from the results presented in Fig. 7, is shown in Fig. 9 Quantitative DNase I footprinting was employed to determine the equilibrium dissociation constant of IHF to its target site on a linearized (relaxed) and a negatively supercoiled DNA plasmid (pDH⌬wt; Table I) 3. In vitro transcriptional analysis of ilvP G promoter activity in the absence and presence of IHF at various superhelical densities. In vitro transcriptions were performed in the presence (ϩIHF; 20 nM) or absence (ϪIHF) of IHF with plasmid DNA template, pDH⌬wt, of various superhelical densities as described under "Materials and Methods." Transcription reactions were terminated after 3 and 6 min and the 157-nucleotide ilvP G 2 (referred to as ilvP G in this work) and 229 nucleotide ilvP G 1 transcripts were isolated by electrophoresis on a denaturing 6% polyacrylamide gel containing 8 M urea and visualized by autoradiography as described under "Materials and Methods." The ilvP G 1 transcripts arises from a fortuitous in vitro promoter in the AϩT-rich UAS1 region (7). Transcription from this promoter is repressed in the presence of IHF (6). The doublet transcription product bands are presumed to arise from heterogeneous termination. visualized by primer extension analysis employing a 32 P-endlabeled 23-bp single-stranded, DNA oligonucleotide that hybridizes to a region immediately upstream of the IHF binding site at ilv bp positions Ϫ155 to Ϫ132 ( Fig. 1; Ref. 16). These DNase I footprinting assays were performed at the same DNA concentration used in the gel mobility shift assays and over a range of IHF concentrations from 0.33 to 330 ϫ 10 Ϫ9 M (Fig. 9). The IHF binding isotherms obtained for its ilvP G target site on the linearized and negatively supercoiled DNA plasmids are shown in Fig. 9. K D (Ϯ S.D.) values of 3.8 Ϯ 1.1 ϫ 10 Ϫ9 M (relaxed) and 3.0 Ϯ 1.5 ϫ 10 Ϫ9 M (supercoiled) based on the free energies of IHF binding of ⌬G ϭ Ϫ11.5 Ϯ 0.2 and ⌬G ϭ Ϫ11.6 Ϯ 0.4 (kcal/mol), respectively, were determined by the methods of Brenowitz et al. (17) as described under "Materials and Methods." The values obtained with DNase I footprinting of IHF binding to relaxed and negatively supercoiled DNA templates are indistinguishable from one another and nearly the same as the value obtained with the gel shift assay. Thus, the DNA supercoiling dependence of IHF-mediated activation of transcription from the ilvP G promoter is not due to differential binding affinities of IHF to relaxed and supercoiled DNA templates. DISCUSSION We have previously demonstrated that IHF binds to the UAS1 region upstream of the ilvP G promoter (8), and wraps the DNA around the body of the protein to form a higher-order nucleoprotein complex (6,8) and facilitates the unwinding of the DNA helix in the Ϫ10 hexanucleotide region of the downstream promoter (7). We showed that IHF affects the transcription initiation reaction at the ilvP G promoter by increasing the rate of open complex formation (7). We further demonstrated that IHF-mediated activation of transcription from this promoter requires a supercoiled DNA template (6), occurs in the absence of specific interactions between IHF or upstream DNA sequences or transcription factors and RNA polymerase (7), is face-of-the-helix and orientation independent (6, 7), and can be replaced by a heterologous DNA bending protein (7). These results led us to exclude previously described activation models that are based on DNA looping, or require specific IHF-RNA polymerase interactions and are dependent on the orientation of the DNA helix in the upstream promoter region (7). Instead, we proposed a novel allosteric DNA mechanism. We suggested that IHF activates transcription from the ilvP G promoter by forming a higher-order protein-DNA complex in the UAS1 region of a supercoiled DNA template that structurally alters the DNA helix in a way that facilitates open complex formation at the downstream promoter site. To further investigate this  model, we wished to examine the requirement of a supercoiled DNA template for IHF-induced activation.
In bacteria the degree of DNA supercoiling is maintained, primarily, by the dynamic balance of two topoisomerases, gyrase (the gene product of the gyrA and gyrB genes) which adds negative DNA supercoils and topoisomerase I (the gene product of the topA gene) which removes DNA supercoils (1-3). Blockage of either activity, either by small molecule inhibitors or by structural gene mutations, is known to affect the transcriptional activities of a large number of promoters (1). In principal, increased negative DNA supercoiling can increase transcrip-tion initiation by increasing the free energy of binding of RNA polymerase to a DNA template or by decreasing the energy of activation required for the isomerization of the RNA polymerase-promoter complex from a closed to an open form (29). Since isomerization is the rate-limiting step of the transcription initiation reaction from the ilvP G promoter (7) and increased DNA supercoiling enhances DNA duplex destabilization (30), it is reasonable to presume that increased negative DNA supercoiling increases transcription from this promoter by increasing the rate of the isomerization. IHF also activates transcription from this promoter by increasing the rate of the isomerization step and, in the presence of IHF, the activating effect of increased negative DNA supercoiling is enhanced. However, although the in vivo DNA supercoiling of the bacterial chromosome is slightly higher in a wild type strain in the presence of IHF than in its absence (Table II; Ref. 31), it is clear that IHF does not activate transcription simply by increasing DNA supercoiling or by binding better to supercoiled DNA. These conclusions are based on the in vitro observations that: (i) although promoter activity is increased by DNA supercoiling, both in the presence and absence of IHF the relative activity of the promoter remains the same over a wide range of template DNA superhelical densities (Fig. 5); and (ii) IHF binds with indistinguishable affinities to relaxed and supercoiled DNA templates ( Fig. 9; Ref. 32). Thus, although both increased DNA supercoiling and IHF facilitate the unwinding of the DNA helix and the rate of promoter open complex formation (7), our results suggest that the activation of this rate-limiting step in the transcription initiation reaction by DNA supercoiling and IHF are effected by distinctly different mechanisms.
An effect of DNA supercoiling on the IHF-mediated activation of another procaryotic promoter has been reported. Higgins et al. (33) have proposed a "superloop" model to explain the DNA supercoiling dependent IHF-mediated activation of transcription from the bacteriophage Mu P E promoter. They suggest that the stabilization of IHF at the apex of a supercoiled DNA loop impedes the twisting of the double helix about its axis and causes a persistent face of the DNA helix to be exposed on the outside face of the loop. They propose that activation is  facilitated at an optimal superhelical density by the positioning of the P E promoter halfway between the IHF binding site and the first supercoil node where it is in an ideal location for interaction with RNA polymerase. In support of this model, van Rijn et al. (34) demonstrated that the activity of the Mu P E promoter is modulated with a 10-bp periodicity as it is incrementally moved farther away from and closer to the IHFmaintained apex of the DNA superloop. This type of a face-ofthe-helix or DNA supercoiling dependence is not observed for IHF-mediated activation of transcription from the ilvP G promoter. Thus, the superloop model can be excluded as a primary explanation for IHF-mediated regulation of transcription from the ilvP G promoter.
Bauer et al. (35) have used computational methods to predict the sites on plasmid DNA molecules where superhelical stresses destabilize the duplex. Duplex destabilization is predicted to occur in these molecules at a small number of AϩTrich sites and these predictions agree well with biochemical measurements (36). In these energetically closed systems an increase in the unwinding at one of these sites by increased DNA supercoiling predicts decreased unwinding at another site. These observations present an interesting model for the IHF-mediated activation of ilvP G promoter activity. For example, at physiological superhelical densities, strand separation in a 43-bp AϩT-rich sequence upstream of and including the IHF binding site in the UAS1 region contained in a negatively supercoiled pBR322 based plasmid is observed. 2 Perhaps IHF binding decreases the probability of duplex destabilization at its binding site in UAS1 and increases the probability of unwinding the DNA helix in the Ϫ10 region at the downstream ilvP G promoter site (7). Such a DNA structural transmission model is capable of explaining the facts that IHF activation of transcription from the ilvP G promoter requires a supercoiled DNA template and occurs in the absence of specific protein interactions without altering the overall superhelical density of the DNA template (Fig. 5).
In conclusion, it is interesting to note that the total intracellular concentration of IHF has been estimated to be very high (1 ϫ 10 Ϫ5 M; Ref. 37). Yang and Nash (32) have determined that the free intracellular concentration of IHF is sufficient to saturate a number of variant IHF binding sites with apparent equilibrium binding affinity constants up to 10 times greater than the K D of IHF for its binding site in the ilvP G promoter region. It is, therefore, likely that this IHF site is always occupied in vivo. Thus, the IHF-DNA complex in UAS1 might be considered a permanent architectural feature of this promoter. If this is the case then the question arises: is IHF merely an architectural promoter element or does it also have a physiologically important regulatory purpose? This becomes a particularly intriguing question when it is considered that no physiological regulatory role for IHF has been discerned for any of the many operons whose expression it is known to affect (12). The results reported here, however, suggest an important physiological role for IHF-mediated regulation of the ilvG-MEDA operon. This suggestion is based on the fact that the global superhelical density of the chromosome varies over a wide range during different phases of the bacterial growth cycle (23) and in response to various types of environmental assaults such as osmotic, temperature, and anaerobic shocks, or nutritional upshifts and downshifts (reviewed in Ref. 1). We show in this report that IHF functions to adjust the basal level of ilvP G transcription as a function of superhelical DNA density. Thus, the physiological role for this global transcriptional activator for the ilvGMEDA operon seems to be to amplify the response of the ilvP G promoter to small changes in superhelical density effected by changes in growth conditions. Such a mechanism might coordinate the capacity for branched chain amino acid biosynthesis with the growth conditions of the cell.