Organization of Open Complexes at Escherichia coli Promoters LOCATION OF PROMOTER DNA SITES CLOSE TO REGION 2.5 OF THE s 70 SUBUNIT OF RNA POLYMERASE*

A cysteine-tethered DNA cleavage agent has been used to locate the position of region 2.5 of s 70 in transcriptionally competent complexes between Escherichia coli RNA polymerase and promoters. In this study we have engineered s 70 to introduce a unique cysteine residue at a number of positions in region 2.5. Mutant proteins were purified, and in each case, the single cysteine residue used as the target for covalent coupling of the DNA cleavage agent p -bromoacetamidobenzyl-EDTA z Fe (FeBABE). RNA polymerase core reconstituted with tagged s derivatives was shown to be transcriptionally active. Hydroxyl radical-based DNA cleavage mediated by tethered FeBABE was observed for each derivative of RNA polymerase in the open complex. Our results show that region 2.5 is in close proximity to promoter DNA just upstream of the 2 10 hexamer. This positioning is independent of promoter sequence. A model for the interaction of this region of s with promoter DNA is discussed.

Promoter recognition requires sequence-specific contacts by the transcriptional apparatus. At most promoters these contacts are made upstream from the transcription start point. Once the transcriptional apparatus has bound the promoter to form a closed complex, an isomerization event occurs to generate the open complex, forming the single-stranded template required for transcription. The bacterium Escherichia coli provides a good model for understanding protein-DNA interactions during transcription initiation. E. coli uses a single core RNA polymerase for transcription elongation with subunit composition ␣ 2 ␤␤Ј. Promoter specificity is principally afforded by a separate subunit, , which associates with the core enzyme to give holoenzyme (RNAP) 1 but dissociates once sequence-specific promoter DNA contacts are no longer required (1). The 70 subunit, encoded by rpoD, is one of several subunits utilized by E. coli and is responsible for directing the transcription of most genes during vegetative growth. RNAP is capable of sequence-specific transcription initiation in the absence of other transcription factors. Factor-independent transcription is reliant on the ability of 70 to make stable contacts with the promoter DNA (1,2). E. coli promoters contain two very conserved motifs, the Ϫ10 and Ϫ35 hexamers (3), and several less-conserved sequences including the UP element (4) and the extended Ϫ10 motif (5). The extended Ϫ10 motif (5Ј-TGXTATAAT-3Ј) can drive factor-independent transcription at several bacterial promoters lacking homology to the consensus within the Ϫ35 region (6,7). Therefore this TGX motif is able to compensate for a poor Ϫ35 hexamer. The TG motif has been shown to be important for promoter activity in several other bacterial species (8 -11). Work from many laboratories has defined the regions within RNA polymerase that are responsible for sequence-specific contacts within promoter DNA. Regions 2.4 and 4.2 of 70 contact the Ϫ10 and Ϫ35 hexamers, respectively (1,2), whereas the C-terminal domain of the ␣ subunit (␣CTD) contacts the UP element (4). Recent work from this laboratory has indicated that a newly defined region of 70 , region 2.5, is responsible for making sequence-specific contacts with the extended Ϫ10 motif (12). This region was identified by screening for altered or relaxed specificity mutants of 70 capable of compensating for down-mutations within the extended Ϫ10 motif. One relaxed specificity mutant, 70 E458G, was isolated (12). The E458G substitution partially suppressed the effect of changing the G⅐C base pair of the TG motif, suggesting a role for the side chain at position 458 in contacting the extended Ϫ10 motif.
The aim of the study presented in this paper is to complement the genetic study with biophysical data to support the suggested role of region 2.5. We wanted to show that, in open, transcriptionally competent complexes at E. coli promoters, region 2.5 of 70 is located near to promoter DNA, just upstream of the Ϫ10 hexamer. To do this, we exploited a novel method that relies on tethering a DNA cleavage agent to a single specific amino acid side chain of a protein (13). The reagent p-bromoacetamidobenzyl-EDTA⅐Fe (FeBABE) has one reactive group facilitating covalent attachment to cysteine side chains, whereas a second group holds a single metal atom in a tight coordination complex (14). Under appropriate conditions, the divalent cation can participate in the generation of hydroxyl radicals, which attack deoxyribose units, resulting in DNA strand scission (15). Recently, this chemistry has been applied to the study of the interaction of ␣CTD of E. coli RNAP with promoter DNA. Hydroxyl radical DNA cleavage mediated by FeBABE showed that the two ␣CTD subunits are arranged asymmetrically, contacting different halves of the UP element and that activator contact patches are available on both subunits (16,17). To identify sites on promoter DNA that are near to region 2.5 of 70 in open complexes, amino acids in this region were replaced with cysteine for conjugation with the DNA cleavage agent FeBABE.

EXPERIMENTAL PROCEDURES
Strains and Materials-E.coli strain DH5␣ (supE44⌬lacU169(80 lacZ⌬M15) hsdR17 recA1 endA1 gyrA96 thi-1 relA1) was used for all * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
ʈ To whom correspondence should be addressed: Tel.: ϩ44-121-414-5438; Fax: ϩ44-121-414-7366; E-mail: s.d.minchin@bham.ac.uk. cloning (18). Unless indicated otherwise, chemicals were purchased from Sigma, radiochemicals were from NEN Life Science Products, synthetic oligonucleotides were from Alta Bioscience at the University of Birmingham, restriction and DNA-modifying enzymes were from New England Biolabs and Taq DNA polymerase from Boehringer Mannheim.
Plasmids Encoding 70 -Mutants of rpoD encoding a single cysteine were generated by megaprimer polymerase chain reaction (19) using plasmid pGEMD(S), which contains a cysteine-free rpoD gene as a template (20,21). Mutagenic oligonucleotides complementary to the template strand of rpoD used were CYS454 (5Ј-CCATCCGTATTCCGTGTCACATGATTGAG-ACCATC-3Ј), CYS459 (5Ј-GTGCATATGATTGAGTGTATCAACAAGCTC-AAC-3Ј), and CYS461 (5Ј-TGATTGAGACCATCTGTAAGCTCAACCGTAT-T-3Ј), encoding mutations corresponding to cysteine substitutions at positions 454, 459, and 461 of 70 . In the first round of polymerase chain reaction, a mutagenic oligonucleotide and oligonucleotide D13346 (5Ј-GGTCGCAGA-ATCCAGCGGC-3Ј), annealing to the coding strand of rpoD were used to amplify a "megaprimer" fragment of rpoD. In the second round of polymerase chain reaction, the megaprimer and oligonucleotide D12444 (5Ј-GCAGATTAATGATATCAACCGTCGT-3Ј) annealing to the template strand of rpoD were used to amplify a 540-base pair fragment of rpoD. Plasmid cloning vector Pinpoint-Xa1 (Promega) was used for direct cloning of polymerase chain reaction products according to the manufacturer's instructions, and the recombinants were used for sequencing. A cysteine substitution at position 458 had previously been constructed (22). Restriction enzymes XhoI and BamHI were used to subclone internal fragments of rpoD containing the desired mutations into pGEMD(S) (21). Construction of the other single cysteine mutants used (422C and 581C) are described elsewhere (21).
Promoter Fragments-The KAB-TTcon (23) and KAB-TG (24) promoters were cloned on EcoRI-HindIII fragments in the galK fusion vector pAA121. The promoter galP1(19T8A9A) (25) was cloned as an EcoRI-HindIII fragment in pBR322. The sequence of the three promoters is shown in Fig. 1. For FeBABE cleavage reactions and transcription run-off assays, 850-base pair PstI-HindIII fragments were prepared from plasmid DNA purified on caesium chloride gradients. In addition to the promoter under study, these fragments also contained two additional promoters, pX and pbla. For the FeBABE cleavage analysis, the template strand was labeled at the HindIII end with [␥-32 P]ATP and T4 polynucleotide kinase or the nontemplate strand was labeled at the HindIII end using [␣-32 P]dATP and the E. coli DNA polymerase Klenow fragment.
Proteins-Single cysteine mutants of 70 were purified by the method described previously for the wild-type protein (20,26), with the important modification of the exclusion of thiol-containing reducing agents (dithiothreitol or ␤-mercaptoethanol) from all steps in purification. Purified mutant subunits were dialyzed against storage buffer (10 mM Tris-HCl, pH 7.6, at 4°C, 10 mM MgCl 2 , 0.1 mM EDTA, 100 mM KCl, and 50% glycerol) and stored at Ϫ20°C.
Conjugation of 70 with FeBABE-FeBABE was synthesized and characterized as described previously (27). Single cysteine derivatives of 70 were used for conjugation with FeBABE based on the method of Murakami et al. (16). Conjugation was initiated by mixing 1.2 ml of 6.67 M protein solution in conjugation buffer (10 mM Hepes, 200 mM KCl, and 2 mM EDTA, pH of 8.0) with 9 l of 18 mM FeBABE in Me 2 SO. After incubation at 37°C for 4 h, excess unreacted FeBABE was removed by dialysis against storage buffer. The efficiency of conjugation was determined by estimating free side chains of both conjugated and un-conjugated proteins with the fluorescent reagent CPM (7-diethylamino-3-(4Јmaleimidylphenyl)-4-methylcoumarin) (Molecular Probes) (27).
Reconstitution of RNA Polymerase Holoenzyme-A 10-fold molar excess of 70 was mixed with core RNAP and incubated at 20°C for 30 min (26).
DNA Cleavage by FeBABE-RNAP holoenzyme (300 nM) was mixed with 32 P-end-labeled promoter fragment (0.4 nM) in a reaction volume of 35 l (20 mM HEPES, pH 8.0, 5 mM MgCl 2 , 50 mM potassium glutamate, 50 g/ml bovine serum albumin, 5 g/ml poly(dI⅐dC)) and incubated at 37°C for 20 min. Complexes were challenged with heparin (200 M at 37°C for 5 min). DNA cleavage was initiated by the addition of sodium ascorbate (2 mM) followed by incubation at 37°C for 20 min. Modified DNA was extracted with phenol/chloroform and precipitated with ethanol before analysis on a 6% polyacrylamide gel containing 6 M urea. Gels were calibrated with Maxam-Gilbert GϩA sequence ladders and were processed and scanned using a PhosphorImager (Molecular Dynamics).
In Vitro Transcription-The activity of RNAP reconstituted with wild-type, un-conjugated, and conjugated 70 derivatives was measured by in vitro transcription assays (28). Fragments used for cleavage analysis were also used as templates for in vitro transcription. The derivatives of the gal promoter were expected to generate run-off products of 51 nucleotides. DNA template (5 nM) and RNAP holoenzyme (100 nM) were preincubated in 12 l of transcription buffer (10 mM Tris, pH 8.0, 50 mM NaCl, 2.5 mM MgCl 2 , 50 M EDTA, 0.5 mM dithiothreitol, 25 g/ml bovine serum albumin, 2.5% glycerol) for 5 min at 37°C. Transcription was initiated by the addition of unlabeled nucleotide triphosphates ATP, CTP, GTP (200 M), and UTP (10 M), 0.5 Ci of [␣-32 P]UTP (800 Ci/mmol), 100 g/ml heparin. Reactions were stopped by the addition of an equal volume of run-off stop mix (20 mM EDTA, 80% deionized formamide, 0.1% bromphenol blue, 0.1% xylene cyanol). Transcripts were analyzed on a 6% polyacrylamide gel containing 6 M urea and scanned using a PhosphorImager.
Modeling of RNA Polymerase-Promoter Interactions-Modeling of region 2.5 interactions with promoter DNA is based on secondary structure prediction and genetic evidence. Secondary structure prediction  suggests that Glu-458 is part of a ␣-helix starting at Val-454. An octapeptide from Val-454 to Asn-461 was constructed using the molecular modeling package Quanta by Molecular Simulations, Inc. and energy minimized to place amino acid side chains in sterically favorable positions. This peptide was then manually docked into the major groove of a model of B-form DNA based on the sequence of KAB-TG from position ϩ1 to Ϫ29 with the carboxyl group of Glu-458 making base-pair edge hydrogen bond interactions with N6 of adenine at position Ϫ15 and N4 of cytosine at position Ϫ14 (i.e. the template strand of the TG motif). In this position, no steric clashes were observed. The model was extended to include the helix of region 2.4 (helix 14) derived from the crystal structure of a tryptic fragment of 70 (29). The atomic coordinates for the fragment (1SIG) were obtained from the Protein Data Bank (Brookhaven National Laboratory, Upton, Long Island 2 ). Genetic studies show that Gln-437 and Thr-440 are involved in interactions with position Ϫ12 (30,31). The coordinates obtained were used to dock helix 14 with residues 437 and 440 in hydrogen-bonding contact with the base pair edge at Ϫ12 (see Fig. 7). The minor groove at the center on the Ϫ10 element must be placed on the inside of a curvature for efficient promoter recognition by E. coli RNAP (32). In addition, many studies show that recognition of Ϫ10 and Ϫ35 elements is accompanied by promoter bending and suggest that the major groove of the Ϫ10 hex-amer widens to accommodate (33). Such a promoter structure would allow favorable interactions between basic residues Arg-441 and Arg-446 and the phosphate backbone (T-A base steps of the Ϫ10 promoter consensus element distort double-strand DNA in solution) (34). The model in Fig. 7a places the carboxyl group of amino acid 448 1.0 nm from the amino group of residue 454. This would allow for a flexible unstructured 7-amino acid loop connecting the two helices shown. Methods of probing for single-stranded DNA assign position Ϫ12 as the upstream limit of the open complex (35, 36). The TG motif at position Ϫ14/15 would thus remain in a region of double-stranded DNA, whereas helix 14 is shown within the transcription bubble (Fig. 7b). 70 Mutants with Fe-BABE-In previous work, we used suppression genetics to identify a region of the 70 subunit of RNA polymerase, region 2.5, that interacts with the extended Ϫ10 motif of bacterial promoters (12). In this work, we have exploited a tethered DNA-cleaving agent to show that region 2.5 of 70 is in close proximity to the extended Ϫ10 motif. The reagent used was FeBABE, which is covalently attached to 70 by conjugation with a cysteine residue. Starting with an rpoD gene that had been mutated to remove all three native cysteine codons, single 2 http://www.pdb.bnl.gov.  With the exception of 422C, all derivatives retained at least 80% activity compared with wild-type before conjugation and at least 70% activity after conjugation with FeBABE. Specificity of FeBABE Cleavage-In our first experiment, we examined the cleavage of a labeled PstI-HindIII fragment purified from pAA121 carrying the KAB-TG promoter (Fig. 1), using RNAP containing 70 tagged with FeBABE positioned at 461. This DNA fragment contains the KAB-TG promoter as well as the pbla and pX (these promoters are located upstream from the EcoRI site in the pAA121 vector). The results in Fig.  4 show that cleavage is observed with holoenzyme reconstituted from FeBABE-conjugated 461C mutant . In contrast, no cleavage was observed with holoenzyme containing either wildtype 70 or unconjugated 461C. The results (Fig. 4, lane 4) clearly show that cleavage is restricted to the three promoters. Further analysis revealed that similar patterns of FeBABEmediated cleavage are observed at the KAB-TG, pbla, and pX promoters (data not shown).

Construction and Conjugation of
In our second set of experiments, both strands of the consensus extended Ϫ10 promoter galP1(19T8A9A) and the semisynthetic promoters KAB-TG and KAB-TTcon ( Fig. 1) were analyzed for cleavage by holoenzymes carrying 70 protein tagged with FeBABE at different positions. The first promoter, galP1(19T8A9A), is a derivative of galP1, which has been changed to introduce a consensus Ϫ10 element. It is an extended Ϫ10 promoter containing a UP element but has a Ϫ35 hexamer with no homology with the consensus (25). To investigate the effects of a Ϫ10 extension on the pattern of cleavage, we also compared the KAB-TG and KAB-TTcon promoters. These promoters, which are also derived from the galP1 sequence, have similar activities. KAB-TG carries an extended Ϫ10 motif and canonical Ϫ10 (4/6 fit to consensus) and Ϫ35 (5/6 fit to consensus) hexamer sequences. KAB-TTcon lacks the extended Ϫ10 motif but carries an improved Ϫ35 (6/6 fit to consensus) hexamer.
Nontemplate Strand Cleavage- Fig. 5a shows  (37). The data here are in agreement with those previously published. FeBABE attached to 581C cleaved promoter DNA at positions Ϫ44/Ϫ45, Ϫ34 to Ϫ37, and Ϫ24 to Ϫ26. FeBABE positioned on 422C cleaved promoter DNA with very low efficiency at position Ϫ13/Ϫ12. The FeBABE positioned within region 2.5 (at residues 454, 458, 459, and 461) resulted in cleavage around Ϫ20 for all the 70 derivatives, with additional cleavages at other positions being dependent on which 70 derivative was studied. The predominant position of cleavage resulting from FeBABE tethered at 458 is ϳϪ20 for all three promoters; however, additional upstream cleavage around Ϫ36 is seen for complexes at the KAB-TG promoter. Cleavage by FeBABE attached to 459C is limited to DNA around position Ϫ20. FeBABE located at 454 cleaves the nontemplate strand at position Ϫ20 but also at Ϫ17/Ϫ16. Cleavage at Ϫ17/Ϫ16 is not observed for other 70 derivatives. FeBABE positioned at 461 again cleaves at Ϫ20, but additionally, there is cleavage from Ϫ13 to Ϫ1.
Template Strand Cleavage- Fig. 5b shows cleavage on the template strand of the different promoters by FeBABE, located at different positions in 70 . Differences in the positions and intensities of DNA cleavage were observed compared with the nontemplate strand. A single region of promoter DNA at positions Ϫ13 and Ϫ14 is sensitive to hydroxyl radical attack by FeBABE when attached to all four positions in region 2.5. In addition FeBABE tethered at 458 and 459 cleaves at positions Ϫ21/Ϫ22, and 458C generates more distal cleavage at Ϫ38 to Ϫ40. FeBABE at 454C also gives unique cleavage at Ϫ18/Ϫ19 and, in common with 461C, cleaves downstream at Ϫ7/Ϫ8.

Region 2.5 of RNA Polymerase 70 Subunit 2267
There is increased cleavage at Ϫ7/Ϫ8 by FeBABE positioned at 461. Faint upstream cleavage for FeBABE attached at 461 was seen at Ϫ21/22 (common to 458C and 459C). Promoter-specific differences were seen in the cleavage pattern from FeBABE at 454, with an increase in the intensity of bands at Ϫ8 corresponding to an alteration of promoter bias toward the Ϫ10 element. As previously observed, FeBABE positioned at 581 in conserved region 4.2 cleaves promoter DNA from Ϫ38 to Ϫ41 and at Ϫ28/Ϫ29, whereas 422C in conserved region 2.3 gives weak cleavage at Ϫ16/Ϫ17. Fig. 6 shows the cleavage pattern from both strands of promoter KAB-TG in schematic form.

DISCUSSION
This study and previous work (37) has shown that the cysteine residues present in the wild-type 70 subunit of RNA polymerase are not essential for transcription initiation. Thus, it is possible to introduce cysteine residues in 70 and to attach the DNA cleavage agent FeBABE and still retain transcriptional competence. In this work, FeBABE attached to the 70 subunit of RNA polymerase was used to probe open complexes at a set of related promoters. The first important point to note is that the overall structure of RNA polymerase bound at promoters carrying different sequence elements appears to be very similar; the cleavage data was similar for all three variants of the galP1 promoter (Fig. 1) as well as the pX and pbla promoters. FeBABE attached to positions 422 in region 2.3 and 581 in region 4.2 were used as controls in our experiments, and the data are similar to those observed in a previous study (37). This is consistent with the view that RNA polymerase uses a variety of protein-DNA interactions to form essentially the same open complex at different classes of promoters. Note that the principal information gained from this technique concerns location, and the relative intensities of different bands cannot be interpreted to give detailed information about binding mechanisms. For this reason, we chose to work with strong promoters, where open complex formation would not be hindered by the bulky substitution of FeBABE.
The major conclusion from this study is that, in open complexes, region 2.5 of 70 is close to promoter DNA sequences just upstream of the Ϫ10 hexamer. One aim of this work was to propose a model for the interaction of region 2.5 of 70 with promoter DNA. Interpretation of the data is complicated by the fact that the holoenzymes carrying FeBABE-modified 70 proteins are interacting with DNA that is known to be both bent and unwound. However, Fig. 7 shows possible models of how regions 2.4 and 2.5 of 70 may interact with promoter DNA, based on the data obtained from this study and previous genetic work as discussed below. The similarity of cleavage pattern observed for FeBABE tethered at 454 and 461 and for FeBABE tethered at 458 and 459 is consistent with region 2.5 being ␣-helical. The cleavage pattern seen for FeBABE at 459 indicates that the reagent contacts the double-stranded promoter DNA around position Ϫ20. In the model, consistent with the length of the FeBABE spacer arm, the ␤ carbon of Thr-459 is approximately 1.2 nm away from the proposed position of radical release. The FeBABE attached to 459C may be constrained, allowing it to sit in only one position relative to the DNA. We propose that position 459 is buried by the helix of region 2.5. FeBABE positioned at 458 results in the same cleavage pattern seen for 459C, but with additional weaker upstream DNA cleavage around the Ϫ35 hexamer, which may be because of bending and wrapping of upstream DNA sequences (33). The cleavage patterns observed with FeBABE positioned at 454 and 461 are more complex. We suggest that this complexity arises from the location of these side chains in an exposed position in the region 2.5 helix and the fact that promoter DNA is melted downstream from position Ϫ12 (12,38).
The model presented in Fig. 7b shows a conformation for open complex promoter DNA consistent with the cleavages observed from FeBABE at positions 454, 458, and 459, and 461. The modeling of region 2.5 as an ␣-helix is consistent with the data where FeBABE at positions 458 and 459 (100°apart on the ␣-helix) only cleave promoter DNA upstream on the TG motif, and FeBABE cleavages from positions 454 and 461 span the TG motif. The FeBABE cleavage upstream of the TG is best modeled with double-stranded DNA. In contrast, the complex pattern downstream of the TG motif cannot be modeled on double-stranded DNA, because the DNA was probed in the open complex. Modeling of region 2.5 as an ␣-helix in contact with the TG motif has important consequences for the orientation of 70 and, in particular, region 2.4. Thus, in addition to the results presented here, we used previous genetic data to orientate both regions 2.4 and 2.5 with respect to the promoter (12,30,31). The aromatic residues Tyr-430 and Trp-433 are implicated in DNA melting and believed to interact at the Ϫ10/Ϫ11 positions on promoter DNA. The Gln-437 and Thr-440 residues interact with the base pair at Ϫ12. The Glu-458 residue is involved in the binding of the extended Ϫ10 motif at Ϫ14/Ϫ15 by region 2.5 of 70 . Hence the structures shown in Fig. 7, a and b, represent proposed orientations of regions 2.4 (helix 14) and 2.5 that can account for both the genetic and the biophysical data. Note that the orientation of helix 14 proposed in Fig. 7 is different to that suggested by Owens et al. (37) on the basis of cleavage patterns generated by FeBABE, located at positions 132, 376, 396, and 422 in 70 . Our present data is insufficient to prove either proposal (because of flexibility both in the DNA and in 70 just downstream of helix 14). For example, increased DNA distortion could fit the data presented by Owens et al. to the models shown in Fig. 7. Similarly our data could be fitted to the Owens et al. model if a sharp kink is introduced into 70 between helix 14 and region 2.5.
According to the model presented here (Fig. 7b), the seven amino acids immediately downstream of helix 14 form a junction between separate domains of 70 (39). Flexibility of the loop would allow movement of region 2.4 relative to 2.5 during open complex formation. In the closed complex, this loop constrains the helix of region 2.4 relative to promoter DNA in the orientation shown (Fig. 7b). This model shows how region 2.5 can serve as an anchor, providing a scaffold on which the open complex may be built. These observations support the idea that the TG motif sets a limit on the conformational fluctuation of the Ϫ10 region (34,40,41). This is consistent with analysis of the temperature dependence of promoter opening at galP1 and supports a mechanism of open complex formation whereby melting nucleates around Ϫ10 (42,43). Such a feature would be of particular importance at extended Ϫ10 promoters that lack an identifiable Ϫ35 hexamer.