Recruitment of RNA Polymerase Is a Rate-limiting Step for the Activation of the ς54 Promoter Pu ofPseudomonas putida *

The activity of the ς54-promoter Pu of Pseudomonas putida was examined in vitro with a DNA template lacking upstream activating sequences, such that RNA polymerase can be activated by the enhancer-binding protein XylR only from solution. Although the transcription activation pathway in this system lacked the step of integration host factor (IHF)-mediated looping of the XylR·DNA complex toward the prebound RNA polymerase, IHF still stimulated promoter activity. The positive effect of IHF became evident not only with XylR from solution, but also with other ς54-dependent activators such as NtrC and NifA. Furthermore, an equivalent outcome was shown for the nonspecific DNA-binding protein HU. This stimulation of transcription in the absence of the enhancer was traced to the recruitment of RNA polymerase (i.e. increased efficiency of formation of closed complexes) brought about by IHF or HU binding. Thus, under limiting concentrations of the polymerase, the factor-mediated binding of the enzyme to Pu seems to enter a kinetic checkpoint in the system that prevents the XylR-mediated formation of an open complex.

The activity of the 54 -promoter Pu of Pseudomonas putida was examined in vitro with a DNA template lacking upstream activating sequences, such that RNA polymerase can be activated by the enhancer-binding protein XylR only from solution. Although the transcription activation pathway in this system lacked the step of integration host factor (IHF)-mediated looping of the XylR⅐DNA complex toward the prebound RNA polymerase, IHF still stimulated promoter activity. The positive effect of IHF became evident not only with XylR from solution, but also with other 54 -dependent activators such as NtrC and NifA. Furthermore, an equivalent outcome was shown for the nonspecific DNA-binding protein HU. This stimulation of transcription in the absence of the enhancer was traced to the recruitment of RNA polymerase (i.e. increased efficiency of formation of closed complexes) brought about by IHF or HU binding. Thus, under limiting concentrations of the polymerase, the factor-mediated binding of the enzyme to Pu seems to enter a kinetic checkpoint in the system that prevents the XylR-mediated formation of an open complex.
Transcription initiation is a sequential multistep process involving promoter DNA recognition by RNA polymerase (RNAP), 1 formation of an initiation-competent RNAP⅐DNA complex, formation of initial phosphodiester bonds, and escape of RNAP from the initial binding site to elongation (1,2). From a kinetic point of view, the overall rate of transcription initiation of a given promoter depends on the slowest phase in the process, so that favoring one nonlimiting step does not result in an increase of the total transcription rate (1,3). Transcriptional activators generally act on these limiting steps to increase promoter output (for review, see Ref. 3). This rule is generally true for the prokaryotic RNAP containing the major sigma factor 70 ( 70 -RNAP). Because positively regulated 70 promoters generally fail to form stable closed complexes (4), activator-mediated binding of 70 -RNAP to cognate promoters is often a limiting step, which, similarly to the eukaryotic counterpart (4,5), is subjected to regulation.
The one exception to this rule is the group of promoters transcribed by the RNA polymerase containing the alternative factor 54 ( 54 -RNAP). In this case, the enzyme is believed to form a stable closed complex with the target DNA sequences at Ϫ12 and Ϫ24 sites (6,7). On the contrary, isomerization to an open complex is strongly stimulated by the action of cognate regulators, generically known as prokaryotic enhancer-binding proteins (8), that bind to upstream activating sequences (UASs) located at Ͼ100 bp from the 54 -RNAP binding site (6). Interactions between 54 -RNAP bound to the Ϫ12/Ϫ24 region and the regulatory protein associated with the UAS are often facilitated by the bending of the intervening DNA by the integration host factor (IHF). IHF is believed to assist the looping out of the region between the RNAP and the activator, thus increasing the overall rate of transcription initiation (9 -13).
Although these notions might be true for most 54 -dependent promoters, we have recently shown that the Pu promoter of the TOL plasmid of Pseudomonas putida (Fig. 1) can barely form a closed complex with its target DNA sequences (14). In this case, the strict dependence of Pu activity on IHF in vivo (15) and in vitro (16) seems to reflect not only the productive geometry of the region brought about by IHF binding but also a more efficient formation of close complexes of 54 -RNAP with the promoter. Such an IHF-mediated "recruitment" of 54 -RNAP seems to involve the interaction of an otherwise distant ciselement with the C-terminal domain of the ␣ subunit of 54 -RNAP (14). This nonanticipated role of IHF was observed in the absence of XylR, the activator of the system, so that the actual effect of IHF-mediated recruitment of 54 -RNAP to Pu on transcription was not substantiated. In this work, we have sought to ascertain this issue by using an in vitro system in which Pu is activated by XylR from solution rather than from the UAS. Our data suggest that 54 -RNAP binding is a ratelimiting step in the process of transcription initiation at the Pu promoter.

EXPERIMENTAL PROCEDURES
Plasmids and General Procedures-All plasmids used in the transcription assays are derived from vector pTE103, which adds a strong T7 terminator downstream of the promoters under study (17). The plasmid called pEZ10 carries the entire region between coordinates Ϫ208 and ϩ93 of the Pu sequence, inserted as an EcoRI-BamHI fragment in pTE103. Plasmid pEZ20 carries the variant named Pu ⌬UAS inserted in the same vector as a 207-bp EcoRI-BamHI fragment excised from plasmid pUC-IHF2 (14), which spans the region Ϫ114 to ϩ93 of Pu. Similarly, a 122-bp fragment from plasmid pUC-d2 (14), containing the region Ϫ53 to ϩ93 of Pu, was cloned in pTE103 to yield plasmid pEZ30, which bears the Pu ⌬UAS ⌬IHF promoter variant. All cloned inserts and DNA fragments were verified through automated DNA sequencing in an Applied Biosystems device. All the supercoiled DNA templates used for in vitro transcription were purified with the Qiagen system. Other recombinant DNA manipulations were carried out as described previously (18).
Proteins and Protein Techniques-Purified factor 54 , NtrC, NtrB, and native core RNAP from Escherichia coli were the kind gift of B. Magasanik. NifA, IHF, and HU proteins were obtained from M. Buck, H. Nash, and T. Baker, respectively. The XylR variant called XylR⌬A is identical to the wild-type protein except for the deletion of its Nterminal module (called the A domain). This variant is fully constitutive and can thus activate transcription from Pu in the absence of any aromatic inducer (16,19). XylR⌬A was purified to apparent homogeneity by metalloaffinity of the His-tagged protein (16).
In Vitro Transcription Assays-Single-round transcription assays were performed as described before (20). Supercoiled DNA templates were used at 5 nM concentration. 50-l reactions were set up at 37°C in a buffer of 50 mM Tris-HCl, pH 7.5, 50 mM KCl, 10 mM MgCl 2 , 0.1 mM bovine serum albumin, 10 mM dithiothreitol, and 1 mM EDTA. Unless indicated otherwise, each DNA template was premixed with 25 nM core RNAP, 100 nM 54 , 25 nM IHF or 75 nM HU, and the concentrations of XylR⌬A, NtrB⅐NtrC, and NifA indicated in each case. Linear DNA templates were generated by digesting the corresponding plasmids (pEZ10, pEZ20, and pEZ30; Fig. 1) with EcoRI, and they were used at the same concentration and conditions as the supercoiled counterparts. The DNA templates and the proteins were incubated at 37°C with 4 mM ATP for 20 min to allow open complex formation. A single cycle of transcription was then initiated by adding a mixture of ATP, CTP, GTP (400 M each), UTP (50 M), [␣-32 P]UTP (5 Ci at 3000/mmol), and heparin (0.1 mg/ml), the latter to prevent reinitiation. After incubating 10 min at 37°C, the reactions were stopped with an equal volume of a solution containing 50 mM EDTA, 350 mM NaCl, and 0.5 mg/ml carrier tRNA. The mRNA extracted and precipitated with ethanol was electrophoresed on a denaturing 7 M urea, 4% acrylamide gel and visualized by autoradiography.
DNase I Footprinting Techniques-DNA-protein interactions were monitored with DNase I footprinting assays performed in a total volume of 50 l of a buffer consisting of 35 mM Tris acetate, 70 mM KAc, 5 mM MgAc 2 , 20 mM NH 4 Ac, 2 mM CaCl 2, 1 mM DTT, 3% glycerol, and 40 g/ml poly[d(I⅐C)]. The DNA template used was a 474-bp BamHI-PvuII fragment excised from plasmid pEZ9 (11), which contains the entire Pu promoter sequence as an EcoRI-BamHI insert in pUC18 spanning positions Ϫ208 to ϩ 93 (Fig. 1). The fragment was end-labeled in its BamHI site by filling in the overhanging end with [␣-32 P]dATP and the Klenow fragment of DNA polymerase. Radioactive nucleotides not incorporated to DNA were removed after a brief spin through small Sephadex G-25 columns. After preincubating the end-labeled fragment (5 nM) for 25 min at 30°C with the proteins indicated in each case, 3 ng of DNase I were added to each sample and further incubated for 3.5 min. Reactions were halted by addition of 25 l of STOP buffer containing 0.1 M EDTA, pH 8, 0.8% SDS, 1.6 M NH 4 Ac, and 300 g/ml sonicated salmon sperm DNA. Nucleic acids were precipitated with 175 l of ethanol, lyophilized, and directly resuspended in denaturing loading buffer (7 M urea, 0.025% bromphenol blue, and 0.025% xylene cyanol in 20 mM Tris, pH 8) before loading on a 7% DNA sequencing gel. AϩG Maxam and Gilbert reactions (21) were carried out with the same fragments and loaded in the gels along with the footprinting samples.

RESULTS AND DISCUSSION
Rationale for Separating Structural Effects of IHF from Recruitment of the 54 -RNAP in the Pu Promoter-IHF protein has been shown to produce two effects on the Pu promoter. On one hand, it provides a structural aid to bring about contacts between the upstream UAS⅐XylR complex and the 54 -RNAP bound to Ϫ12/Ϫ24 (11,13). On the other hand, it augments the affinity of 54 -RNAP for the promoter (14). As a consequence, the observed stimulatory effect of IHF in Pu activity (11,13) should originate from both the optimization of promoter geometry and the increased efficiency of formation of closed complexes. To separate these two effects, we produced a variant of the Pu promoter in which UAS DNA was deleted up to the Ϫ114 site (Pu-114; Fig. 1). Transcription from such a promoter is predicted to miss the step of looping out of the intervening sequence and to rely only on the direct contact between the activator from solution and the 54 -RNAP bound to the Ϫ12/ Ϫ24 site. Thus, we set out to compare Pu-114 activation both in the absence and in the presence of IHF in single-round transcription assays with either the intact promoter region (Pu) or a Pu variant deleted of both the UAS and the IHF site (Pu-53).
To avoid the addition of an aromatic inducer (e.g. toluene) to the in vitro assays, these templates were added with XylR⌬A, a constitutively active form of XylR that is deleted of its Nterminal module (the so-called A domain; Ref. 16). We also predicted that XylR⌬A could activate transcription from the templates deleted of UAS at a higher protein concentration than full-length Pu, as has been observed for 54 -RNAP activation from solution in other 54 -dependent regulators (12,(22)(23)(24)(25). Under these conditions, any effect of IHF in transcription must reflect exclusively the efficiency of formation of closed complexes, because any geometrical effect to bring about XylR-54 -RNA contacts is ruled out.
IHF Stimulates Activation of 54 -RNA by XylR⌬A from Solution-To ascertain whether the increased binding of 54 RNA to Pu caused by IHF (14) was in fact translated into a higher transcriptional rate, we ran in vitro assays with supercoiled plasmids bearing wild type Pu, Pu ⌬UAS (Pu-114), or Pu ⌬UAS ⌬IHF (Pu-53). These templates were incubated with subsaturating concentrations of 54 -RNAP and IHF, along with XylR⌬A, the latter in a 10-fold excess when using templates devoid of the UAS. As expected (16), transcription in any of the conditions tested was absolutely dependent on the presence of the XylR⌬A protein (data not shown), a common feature of all 54 -dependent activators known so far (6,7). Because assays were carried out in the presence of heparin to prevent reinitiation, the transcripts originated from single rounds, and their levels were proportional to the amount of the open complexes formed under different conditions. As shown in Fig. 2A, Pu ⌬UAS could be efficiently transcribed in the presence of XylR⌬A (16) by simply increasing approximately 10-fold the amount of the activator added to the assays compared with the wild-type Pu template. In addition, it became evident that IHF maintained a strong stimulatory effect on transcription of Pu ⌬UAS, not unlike that observed with the complete Pu promoter. This effect was entirely dependent on IHF bound to its site within the Ϫ29/Ϫ114 region, as indicated by the control experiment with the Pu ⌬UAS ⌬IHF template, which lacked any stimulation by the factor (Fig. 2A). The Pu ⌬UAS ⌬IHF DNA was, in fact, a poor template for transcription, most likely because of the loss of the UP-like element, which overlaps the IHF-binding sequence (Ref. 14 and Fig. 1). That the increased The scheme at the top shows the distribution of the functional ciselements of the wild-type Pu segment (coordinates Ϫ208 to ϩ93) included in plasmid pEZ10 with respect to the transcription start site. These include the sequence recognized by 54 -RNAP (Ϫ12/Ϫ24 motif), the binding site for the IHF, and the UASs, which are the targets of the activator of the system, XylR. The location of an UP-like sequence overlapping part of the IHF site and extending further upstream (14) is also indicated. In addition, the vector pTE103 places a T7 terminator (T) downstream of the promoter, so transcripts originated at Pu and its derivatives are 394 nucleotides in size. The bottom schemes show the Pu variants inserted also in pTE103 and used in this study as transcription templates along with the names of the corresponding plasmids. Their inserts span positions Ϫ114 to ϩ22 (Pu ⌬UAS) and Ϫ53 to ϩ22 (Pu ⌬UAS ⌬IHF), respectively. The sequence around the IHF site (Ϫ114 to Ϫ53) is shown for reference. activation of Pu⌬UAS with IHF was not caused by nonspecific binding of XylR⌬A to DNA upstream of the Ϫ114 site in the supercoiled template (Fig. 1) was verified by the experiment shown in Fig. 2B. In this case, linear templates entirely deleted of any sequence upstream of Ϫ208 (wild-type Pu), Ϫ114 (Pu ⌬UAS), or Ϫ53 (Pu ⌬UAS ⌬IHF) were passed through the same transcription assays than the supercoiled counterparts. The data of Fig. 2B show that although Pu ⌬UAS could be stimulated by IHF, the Pu ⌬UAS ⌬IHF template could not. Although the ability of XylR⌬A to activate Pu from solution is reminiscent of that observed in NtrC (12) and NifA (22); such an activation was prevented by the lack of IHF or deletion of the binding site for the factor. The data of Fig. 2 thus strongly suggested that the interaction of 54 -RNAP with Pu limited transcription initiation and that the previously described IHFmediated recruitment of 54 -RNAP (14) could relieve this limitation.
IHF Facilitates Activation of Pu by Other Enhancer-binding Proteins-To ensure that the stimulatory effect of IHF on Pu activation from solution was not restricted only to XylR⌬A, we also assayed two proteins of the family of enhancer binding factors, NtrC and NifA (26,8), known to activate, respectively, the glnHp2 and PnifH promoters from solution (12,22). Because the wild-type Pu does not have binding sites for NtrC or NifA, the assays were made using the complete promoter rather than the version lacking the UAS (27). To this end, purified NtrC and NifA were mixed separately with the Pu template and added or not with IHF before running singleround transcription assays. The reaction with NtrC was amended with purified NtrB protein, which is needed for the activation of NtrC by phosphorylation (28). It was also required to add twice as much of NtrC and NifA to the assays than it was of XylR⌬A, perhaps reflecting some difference in the intrinsic activities of the regulators. In any case, as shown in Fig. 3, the presence of IHF was necessary to produce significant amounts of open complexes with any of the proteins tested. These results provided further evidence that IHF stimulation of open complex formation was independent of the UAS and could be traced to an increased occupation of the promoter by 54 -RNAP.
Promoter Occupation by 54 -RNAP Limits Pu Activation from Solution-The data above indicated that IHF stimulates transcription initiation from Pu even in conditions in which looping effects between 54 -RNAP and XylR⌬A bound to distant sites are ruled out. Because IHF allows the Pu promoter to be occupied at lower concentrations of the polymerase (14), the mechanism for such an activation could imply an increased binding of the enzyme and a subsequent increase in the stability of the closed complexes. The prediction is then that an excess of 54 -RNAP concentration should bypass the need of IHF for full transcriptional activity. To test this issue, we carried out in vitro transcription assays in which the Pu ⌬UAS promoter was mixed with growing concentrations of 54 -RNAP and activated from solution by XylR⌬A in the absence or in the presence of IHF. As shown in Fig. 4, the amount of open complexes in the absence of IHF increased with the concentration of 54 -RNAP added, such that they appeared to be limited only by the occupation of the promoter by the enzyme. As shown in Fig. 4 also, IHF addition did overcome such a limitation, because the system became saturated at lower 54 -RNAP concentrations than without the factor.
HU Enhances Activation of the Pu Promoter in trans by XylR⌬A-Although the data presented above seems to substantiate that IHF increases the binding 54 -RNAP to the Pu promoter, the mechanism might not be trivial. Increasing formation of a closed complex may be the result of protein-protein interactions between IHF and 54 -RNAP. Alternatively, recruitment may result from the change of DNA geometry caused by IHF binding, so that an otherwise distant UP-like sequence is brought into the proximity of the Ϫ12/Ϫ24 motif (14). To discriminate between these two possibilities, we used the activation-from-solution assay described above using HU rather than IHF to examine any potential stimulatory effect. HU has been shown to replace IHF in a variety of assays involving DNA bending (29,30,31). Therefore, if IHF-mediated recruitment of 54 -RNAP were caused by specific protein-protein interactions between the factor and the C-terminal domain of the ␣ subunit of 54 -RNAP, then HU could not replace IHF for the stimulatory effect. On the contrary, if the main effect of IHF were caused exclusively by the indirect structural outcome of binding to the promoter region, then HU could substitute functionally its positive influence. To bring these possibilities into a test, the activities of wild-type Pu and Pu ⌬UAS were compared under various combinations of IHF and HU with an excess of Samples were processed as explained in under "Experimental Procedures." Note the effect of IHF addition in Pu and Pu ⌬UAS, and the lack of any significant activity of Pu ⌬UAS ⌬IHF. B, Linear DNA templates. Transcription reactions were set up and run identically as before but using as templates pEZ10, pEZ20, and pEZ30 linearized upon digestion with EcoRI. This cleaved the plasmids at sites Ϫ208, Ϫ114, and Ϫ53, respectively, and thus entirely deleted the upstream DNA sequences. The concentration of XylR⌬A was increased to 0.5 M in the control assay with wild type Pu to compensate for the loss of affinity of the regulator for relaxed UAS DNA (16). Under these conditions, the effect of IHF on wild-type Pu was less pronounced than with the supercoiled counterpart. No transcripts were detected in the absence of XylR⌬A in any of the conditions tested (data not shown).

FIG. 3. Activation of the Pu promoter by NtrC or NifA proteins in the presence of IHF.
Single-round transcription reactions contained 5 nM supercoiled plasmid pEZ10, which bears the wild-type Pu promoter. This was mixed with 25 nM core RNAP, 100 nM 54 , and, where indicated (ϩ), 25 nM IHF. Purified XylR⌬A was entered in the control sample at a concentration of 100 nM, whereas NifA was added at 200 nM. In the case of NtrC, the protein at 200 nM was combined with a 15 nM concentration of its partner kinase NtrB to phosphorylate the regulator in the presence of ATP (28). Note in all cases the positive effect of IHF addition.
XylR⌬A. As shown in Fig. 5, HU indeed had a positive effect on the activation of Pu by XylR⌬A in trans, albeit less pronounced than IHF. Similar also to the results of Fig. 2, HU had no effect on the transcription of a DNA template deleted of the region upstream of Ϫ53 (data not shown), suggesting that, like IHF, its stimulatory effect required the presence of the UP-like element (Fig. 1). Simultaneous addition of the two factors did not appear to further increase the degree of stimulation achieved with IHF alone. These data support the notion that the recruitment of the polymerase brought about by IHF is caused by indirect structural effects (i.e. approaching an otherwise distant UP-like element), and that protein-protein interactions may not play a significant role.
HU Promotes Occupation of Pu by 54 -RNAP-The notion that HU produces the same effect as IHF on Pu regarding the recruitment of the polymerase was tested directly with a DNase I footprinting assay. To this end, a DNA fragment bearing the entire Pu was mixed with subsaturating concentrations of 54 -RNAP holoenzyme and either purified IHF or HU proteins. The results in Fig. 6 show that the same effect of IHF in promoting 54 -RNAP binding to Ϫ12/Ϫ24 (as revealed by the protection of the sequence from DNase I digestion) could also be achieved by HU. Interestingly, because HU does not interact with an specific DNA sequence but rather promotes the flexibilization of the sequence through transient contacts with the minor groove (32), the recruitment of the enzyme becomes evident without an occupation of the upstream IHF site. Interestingly, the distinct pattern of protected and overdigested bands observed in the region upstream and adjacent to the Ϫ24/Ϫ12 sequence remains the same. This suggests that the same interactions of the 54 -RNAP with the upstream region operatively designated a UP-like element (Ref. 14 and Fig. 1) are facilitated equally well by either of the two proteins. These results favor the notion that it is the structural effect of IHF binding to Pu and not the contacts between the proteins that causes the observed increase in 54 -RNAP affinity and the resulting stabilization of the closed complexes.

Recruitment of 54 -RNAP Is a Rate-limiting
Step for Pu Activation-The changes in DNA conformation required for assembling an orderly promoter geometry represent a kinetic barrier for transcription initiation and may constitute a ratelimiting step of the whole process (3). This notion is exacerbated in 54 promoters, because their activity is dependent on the shape of the DNA segment encompassing the enhancer and the RNAP binding site (6,33). Despite this, isomerization of the  6. DNase I footprinting of the Pu promoter with purified 54 -RNAP, HU, and IHF proteins. The DNA template used was a 474-bp BamHI-PvuII fragment from plasmid pEZ10 containing the entire Pu promoter and labeled with 32 P at its BamHI end. The proteins were added to the samples as indicated at the above the gels at the following concentrations: HU, 50 and 100 nM; IHF, 100 nM; and polymerase, 15 nM core enzyme/50 nM 54 . The AϩG Maxam and Gilbert reaction of the same fragment was used as a reference. The locations of the IHF binding site, the Ϫ12/Ϫ24 motif, and the transcription start site (ϩ1) are indicated to the right.

FIG. 7.
Steps controlling transcription rate of Pu. The scheme pictures how IHF and recruitment of 54 -RNAP may become the ratelimiting step for the activation of the 54 -Pu promoter. The shape and volume of the different proteins is symbolic. From our data it appears that the promoter geometry caused by IHF binding to DNA and the ensuing bending may favor the proximity of the UP-like element to C-terminal domain of the ␣ subunit of 54 -RNAP and perhaps also increase the strength of the contacts (14). In the absence of such a UP-like element (as is the case with Pu ⌬UAS ⌬IHF), the polymerase does not form a closed complex spontaneously; hence the promoter remains inactive. The sole presence of the IHF site and the resulting DNA bending stimulate the recruitment of the enzyme to Ϫ12/Ϫ24, allowing the polymerase to be activated by XylR from solution (Pu ⌬UAS). Such an activation is further increased in the wild-type Pu promoter by virtue of the structural effect, which brings the upstream XylR⅐UAS complex into close proximity to the already bound enzyme. closed 54 -RNAP⅐DNA complex to an open complex has been generally considerered the key bottleneck to be overcome by the cognate activators (6). Once such a barrier is defeated, the transcriptional output depends on the probability of contacts between the activator and the 54 -RNAP bound at distant sites, which, in turn, depends on the intrinsic or protein-induced bending or flexibility of the DNA region involved. The stimulatory effect of IHF in 54 promoters has been interpreted in this context to overcome the hurdle corresponding to this phase. But apart from these geometrical effects, we have observed that the binding of IHF to the Pu promoter also favors the binding of 54 -RNAP to its target sequences at Ϫ12/Ϫ24 (Ref. 14 and Fig. 7). On top of this, we have shown now that polymerase binding becomes a rate-limiting checkpoint in the process of Pu activation. All our data indicate consistently that IHF-mediated recruitment of 54 -RNAP controls Pu output. On this basis, we conclude that formation of a stable closed complex in Pu represents a kinetic barrier that, in cases of limiting concentrations of enzyme, becomes more important than the XylR⌬A-mediated formation of an open complex. This could be effective under physiological conditions (e.g. during the onset of stationary phase) in which the various sigmas compete for a scarce intracellular concentration of core RNAP (34). In this respect, the data of Fig. 4 show that IHF addition and the ensuing recruitment of the enzyme to Pu lowers the concentration of the polymerase required for activation. HU protein appeared to both enhance the recruitment of 54 -RNAP and stimulate Pu transcription in a ⌬UAS promoter, hence reproducing the same stimulatory effect than IHF. This suggests that formation of closed complexes is stimulated by factorinduced changes on the conformation of the DNA, perhaps with little need of protein-protein contacts. It thus appears that although IHF and the C-terminal domain of the ␣ subunit of 54 -RNAP may bind very close or even have overlapping sites in Pu (14), the two proteins may not physically contact, or, even if they do, such contacts appear to be irrelevant for 54 -RNAP recruitment.