Nucleotide-dependent interactions between a fork junction–RNA polymera
http://www.100md.com
《核酸研究医学期刊》
Department of Biological Sciences, Sir Alexander Fleming Building, Imperial College London, South Kensington Campus, London SW7 2AZ, UK
* To whom correspondence should be addressed. Tel: +44 207 594 5442; Fax: +44 207 594 5419; Email: m.buck@imperial.ac.uk
ABSTRACT
Enhancer-dependent transcriptional activators that act upon the 54 bacterial RNA polymerase holoenzyme belong to the extensive AAA+ superfamily of mechanochemical ATPases. Formation and collapse of the transition state for ATP hydrolysis engenders direct interactions between AAA+ activators and the 54 factor, required for RNA polymerase isomerization. A DNA fork junction structure present within closed complexes serves as a nucleation point for the DNA melting seen in open promoter complexes and restricts spontaneous activator-independent RNA polymerase isomerization. We now provide physical evidence showing that the ADP·AlFx bound form of the AAA+ domain of the transcriptional activator protein PspF changes interactions between 54-RNA polymerase and a DNA fork junction structure present in the closed promoter complex. The results suggest that one functional state of the nucleotide-bound activator serves to alter DNA binding by 54 and 54-RNA polymerase and appears to drive events that precede DNA opening. Clear evidence for a DNA-interacting activity in the AAA+ domain of PspF was obtained, suggesting that PspF may make a direct contact to the DNA component of a basal promoter complex to promote changes in 54-RNA polymerase–DNA interactions that favour open complex formation. We also provide evidence for two distinct closed promoter complexes with differing stabilities.
INTRODUCTION
The interactions that RNA polymerases make with transcriptional control proteins and how these proteins interface with signalling pathways is central to achieving appropriate patterns of gene expression. The multisubunit DNA-dependent RNA polymerases (RNAPs) of bacteria and eukaryotes are closely related in structure, having a common line of evolutionary descent (1,2). In bacteria, core RNAP (2??'; E) associates with one of several sigma () factors to form functional holoenzymes (2??'; E) capable of recognizing specific promoter DNA targets (3,4).
Many bacteria contain the enhancer-dependent 54 factor (5–7), and the RNA polymerase holoenzyme containing 54 (E54) has the distinctive property of remaining bound to a promoter in a closed complex that very rarely spontaneously isomerizes to an open complex (6,7). A network of interactions among 54, core RNAP and the promoter DNA are required to stabilize the closed promoter complex conformation (8–10) and hence limit isomerization and formation of open complexes prior to activation (11–13). A DNA fork junction persists within the E54 closed complex, and interactions of 54 with this DNA structure are thought to contribute to maintaining the closed promoter complex (8,11,14,15). Weak binding of holoenzyme containing the 70 (E70) type factor to a fork junction structure is believed to be related to the ability of the E70 to transcribe without the strict requirement of an activator protein (11). Current RNA polymerase promoter complex structures (2,16,17) suggest that RNAP interactions at the fork junction will contribute to delivering the promoter DNA to be melted into the DNA-binding cleft of the core RNAP enzyme.
The use of fork junction containing promoter DNA fragments has greatly enabled structural (16,17) and functional (8,9,18) studies on transcription initiation. Direct recognition of the DNA fork junction structure formed around position –12 with respect to the transcription start site (+1) by 70 and 54 occurs as part of the process of promoter melting. For both classes of factor, protein and DNA isomerization events occur that lead to the formation of open complexes (11,18,19). Isomerization of the basal 54–DNA complex requires ATP hydrolysis by mechanochemical enhancer-binding transcription activator proteins, which changes its interaction with the fork junction DNA and leads to local DNA denaturation (19,20). Activators of E54 belong to the large AAA+ (ATPase Associated Activities) protein family that use ATP binding and hydrolysis to remodel their substrates (19,21,22). Certain DNA fork junction sequences diminish DNA binding by 54 and play a role in restricting isomerization of the closed promoter complex (23,24). The N-terminal Region I of 54 has been shown to be important for tight binding to DNA fork junction structures (8,11,13). The activators bind to the Region I of 54 and use ATP hydrolysis to change fork junction structure interactions made by 54 (19,25). The stable binding of 54 to its activators can be detected in the presence of ADP·AlFx, a transition state analogue of ATP thought to change the functional state of the activator leading to its stable engagement with 54 (25).
Nucleotide-dependent putative intermediates in the 54 transcription activation pathway have recently been identified (26). We now show that ADP·AlFx-dependent binding of the ATP hydrolysing AAA+ domain of the Escherichia coli phage shock protein PspF to E54 changes interactions between the E54 and the fork junction DNA structure present within the closed promoter complex. Interactions between the fork junction DNA and E54 are clearly altered when activator adopts a functional state induced by its interaction with ATP at the point of hydrolysis. A DNA-binding activity within the AAA+ domain of PspF was also evident, indicating that the activator not only makes a direct contact to the 54 protein but may also make contact with a DNA component of a promoter complex. This view is substantiated by the observation that DNA alone stimulates the ATPase activity of the PspF AAA+ domain, where the PspF sequence-specific DNA-binding domain has been deleted. Therefore, 54-mediated interactions between the PspF AAA+ domain and promoter DNA are likely to be of functional importance, and distinct from the tethering of PspF via a DNA-specific binding domain to upstream promoter sequences. Overall the results provide evidence that activation of E54 proceeds sequentially through activator using nucleotide-dependent interactions to remodel a fork junction–E54 complex that then leads to stable DNA opening.
MATERIALS AND METHODS
Proteins
The Klebsiella pneumoniae 54 protein (amino acids 1–477) was prepared as described (23). C-terminal deleted forms of activator PspF lacking a functional DNA-binding domain, PspFHTH (amino acids 1–292) and PspF1–275 (the minimal PspF AAA domain) obtained by overproducing an N-terminal histidine-tagged fragment were used in activation assays (27,28).
The N-terminal histidine-tagged DctD AAA+ domain was overexpressed in E.coli strain ‘Rosetta’ (Novagen) from plasmid pHX187 (a gift from T. Hoover). Freshly transformed cells containing pHX187 were added directly from the agar plate into one litre of ‘Hyper’ media (Athena Enzyme Systems) and grown by shaking at 37°C with good aeration until an OD of 0.7 before induction with 1 mM isopropyl-?-D-galactopyranoside for 3 h. Cells were harvested by centrifugation, resuspended in Buffer A (20 mM Tris, pH 8.0, 65 mM potassium thiocyanate and 5% glycerol) containing ‘Complete’ protease inhibitor cocktail (Roche) and lysed in a cell disrupter at 25 KPSI. The cell lysate was clarified by centrifugation at 15 000 r.p.m. for 30 min at 4°C. The resulting supernatant was applied to a HiTrap 5 ml Nickel chelating column and the His-tagged DctD AAA+ domain eluted from the column by running a 40 ml gradient between 5 and 100% Buffer B (Buffer A containing 1 M imidazole). Column fractions containing the His-tagged DctD AAA+ domain were dialysed overnight into 10 mM Tris, pH 8.0, 50% glycerol, 0.1 mM EDTA, 1 mM DTT and 50 mM NaCl at 4°C, protein concentration measured and protein stored at –80°C.
E.coli core RNA polymerase (E) was purchased from Epicentre Technologies Corp.
DNA duplex formation
Wild-type homoduplex or mismatched Sinorhizobium meliloti nifH and E.coli glnHp2 promoter DNA heteroduplexes were prepared by annealing 88mer oligonucleotides . Pairs of oligonucleotides with one strand 5'-32P-end-labelled and with the unlabelled strand at a 2-fold molar excess were heated at 95°C for 3 min in 10 mM Tris–HCl (pH 8.0), 10 mM MgCl2 and then chilled rapidly in iced H2O for 5 min to allow duplex formation.
Table 1. S.meliloti nifH and E.coli glnHp2 promoter duplexes
Native gel mobility shift assays
Gel mobility shift assays were employed to detect 54 and E54 binary and ternary complexes formed with radioactively labelled S.meliloti nifH and E.coli glnHp2 duplex promoter fragments. 54 (1 μM) or E54 were incubated with 32P-end-labelled duplex DNA (100 nM) at 30°C for 5 min in STA buffer . When required activator (2–20 μM) was added for a further 10 min in the absence or presence (2–4 mM) of a nucleoside triphosphate (GTP), or ADP·AlFx formed in situ by the addition of 0.2 mM AlCl3 to a mixture containing 0.2 mM ADP and 5.0 mM sodium fluoride (25). Unless stated otherwise, non-specific competitor DNA (salmon sperm DNA, 100 ng/μl) was added to binding reactions (see figure legends). A glycerol bromophenol blue loading dye (final concentration 10% glycerol) was added to the samples which were then loaded onto 4.5% native polyacrylamide gels and run at 60 V for 80 min at room temperature in 25 mM Tris and 200 mM glycine buffer (pH 8.6). Unbound DNA and protein–DNA complexes were detected by autoradiography.
DNA footprinting assays
Binding reactions (10 μl) containing 100 nM of 32P-end-labelled duplex DNA were conducted as described above. Footprinting reagents were then added; reactions terminated as indicated below; and bound and unbound DNAs were separated on native polyacrylamide gels. DNA was then excised and eluted overnight. For ortho-copper phenanthroline (o-CuP) footprints 0.5 μl of 0.45 mM CuSO4/2.0 mM 1,10-phenanthroline monohydrate and 0.5 μl of 116 mM 2-mercaptopropionic acid (to start the reaction) were added to the binding assay. After 3 min, the reaction was terminated by quenching with 1.0 μl of 28 mM 2,9-dimethyl-1,10-phenanthroline. For dimethylsulphate (DMS) footprints, 1.0 μl of 5% (v/v) DMS was added to the binding reaction for 2 min after which time the reaction was terminated by adding 50 mM 2-mercaptoethanol. Gel-isolated DNA was eluted into H2O overnight at 37°C. DMS-treated DNA was then cleaved with piperidine (10% v/v) at 90°C for 20 min and the DNA pellet washed with H2O. Recovery of isolated DNA was determined by dry Cerenkov counting, and equal numbers of counts were loaded onto denaturing 10% polyacrylamide gels.
DNA cross-linking assays
Promoter probes were based on the heteroduplex S.meliloti nifH promoter (see Table 1) with a photoactivatable cross link (azidophenacyl bromide) tethered to the DNA via a uniquely positioned phosphorothioate residue (29) either upstream or downstream of the region of heteroduplex (i.e. between bases, –15/–14, –14/–13, –13/–12, –12/–11, –11/–10 and –7/–6) with the phosphorothiolated strand 32P-end-labelled (30). Binding reactions (10 μl) containing 100 nM of 32P-end-labelled cross-linking DNA probe were conducted as described above. Reactions were then irradiated (wavelength of 365 nm) for 5 min (UV Stratalinker 1800; Stratagene) prior to loading onto 4.5% native polyacrylamide gels. Stable complexes and unbound DNA from the native gel were eluted into 1x Laemmli sample buffer (Sigma) for 1 h at 37°C and then further analysed by 12.5% SDS–PAGE.
ATPase activity assays
ATPase activity assays of PspF1–275 in the absence and presence of DNA were carried out as described in (31). Reactions containing 0.8 μM PspF1–275 or 0.8 μM PspF1–275 and 10 μM E.coli glnHp2 probes (see Table 1) or 0.45 μg/μl salmon sperm DNA were started by adding ATP to a final concentration of 4 mM containing 0.06 μCi/μl . Samples were incubated at 30°C and time-points were taken as indicated (see Figure 8). Reactions were stopped by adding 5 vol of 2 M formic acid. Released ADP was separated from ATP by thin layer chromatography and amounts of radiolabelled ADP and ATP measured by phosphorimaging (Fuji Bas-1500, Tina 2.10g software).
Figure 8. Time-course showing ATP hydrolysis activity of PspF1–275 in the absence or presence of DNA. Homoduplex or heteroduplex E.coli GlnHp2 promoter probes, (see Table 1; diamonds or open circles, respectively), or salmon sperm DNA (triangles) were incubated with PspF1–275 prior to starting reactions by adding ATP (see Materials and Methods) and compared to PspF1–275 in the absence of DNA (squares). Homoduplex E.coli GlnHp2 promoter probe alone (stars) did not show ATP hydrolysis activity.
RESULTS
Activator bound to ADP·AlFx changes interactions between the DNA fork junction and E54
We used the minor groove-specific DNA footprinting agent o-CuP to probe binary complexes formed between E54 and homoduplex S.meliloti nifH promoter DNA (Table 1), and the ternary E54 complexes formed with the AAA+ activator PspF lacking its C-terminal DNA-binding domain (PspFHTH, amino acids 1–292) (14,25). To form a stable ternary complex between E54 and PspFHTH we used ADP·AlFx, a transition state analogue of ATP that binds to activators of E54 and allows stable binding of PspFHTH to the 54 protein (25). We wanted to learn if E54–DNA interactions within the closed promoter complex were changed by stable binding of the activator. Any changes would provide evidence for activator promoting changes in DNA interactions that could be associated with the nucleotide-dependent stimulation of open complex formation, but which could occur prior to ADP or Pi release. Complexes were formed between the S.meliloti nifH homoduplex promoter probe and E54 in the absence or presence of PspFHTH and ADP·AlFx. DNA footprinting reactions were then carried out and the DNA complexes resolved on native polyacrylamide gels. Subsequently, the DNA cleavage patterns were examined following denaturing gel electrophoresis of the promoter DNA recovered from native gels. Figure 1A shows a native gel indicating the position of binary E54 complexes and the ternary E54–PspFHTH:ADP·AlFx complex from which o-CuP-treated S.meliloti nifH promoter DNA (with end-labelled template strand) was isolated. Figure 1B shows the DNA cutting patterns from the isolated complexes and free DNA seen in Figure 1A. Compared to promoter DNA not exposed to the E54, the E54-bound DNA shows a clear protection from cutting across the promoter region and a strong extra reactivity just downstream of the GC consensus promoter sequence . The extra reactivity (at position –11) corresponds to the position of the native DNA fork junction structure (8,14). In ternary complexes that contain both E54 and PspFHTH:ADP·AlFx the extra reactivity to o-CuP at position –11 is not seen, but it is present in those binary promoter complexes that contain E54 and which were isolated from the same reaction in which the PspFHTH and ADP·AlFx were present (Figure 1B, compare lanes 5b with 5a). These unstable complexes are potentially unstable binary complexes that PspFHTH is unable to bind or modify. Thus, changed cutting patterns are clearly associated with the stable ternary complex.
Figure 1. Changed promoter interactions within ternary E54–PspFHTH:ADP·AlFx promoter complexes. (A) Native gel showing the position of binary E54 complexes (a) and the ternary E54 complex (b) formed by the addition of PspFHTH:ADP·AlFx using the S.meliloti nifH homoduplex promoter probe (see Table 1) following treatment with the footprinting reagent o-CuP, (B) template strand, and (C) non-template strand, o-CuP footprints from unbound DNA or bound promoter complexes (a) and (b) shown in (A). In (B), the position of the native DNA fork junction structure (at –11) and extra o-CuP reactivity at positions –14 and –13 are indicated by an arrow and unfilled arrowheads, respectively. Reactions contain 100 nM promoter probe, 250 nM E54 , 20 μM PspFHTH and 100 ng/μl salmon sperm DNA.
Using homoduplex S.meliloti nifH promoter DNA with the non-template strand end-labelled, complementary results were obtained (Figure 1C). Again the extra fork junction reactivity towards o-CuP seen in DNA bound by the E54 was absent when the PspFHTH:ADP·AlFx was bound to the E54 (Figure 1C, compare lanes 5a and 5b respectively). Controls showed that ADP·AlFx in the absence of PspFHTH did not change the patterns of o-CuP reactivity across the E54-binding region (Figure 1B and C, lane 6a) and that PspFHTH alone did not change the E54 footprint (Figure 1B and C, lane 4a). Further, core RNAP (E) did not yield protection or extra cutting regardless of the presence of ADP·AlFx (Figure 1B and C, compare lanes 7 and 8). Since the footprinting reactions contained salmon sperm DNA as non-specific competitor DNA and PspFHTH:ADP·AlFx alone did not cause changes in cutting patterns under these conditions (Figure 1B and C, compare lanes 2 and 9), we conclude that the PspFHTH:ADP·AlFx causes an altered interaction of o-CuP with promoter DNA in ternary E54 complexes at the place where the fork junction structure is maintained. Attempts to probe the conformation of the A:T base-pair 3' to the 5' GC sequence at –12 using KMnO4 or diethylpyrocarbonate failed because complexes containing PspFHTH:ADP·AlFx were largely destroyed under reaction conditions. Clearly the presence of the activator with ADP·AlFx bound causes changes, in o-CuP reactivity within the closed complex, near the GC promoter element, but does not distinguish between the PspFHTH:ADP·AlFx causing protection of the –12 proximal sequences from cutting or causing a change in the structure of DNA in the –12 region to account for the absence of hyper-reactivity in the ternary complexes.
Further examination of the DNA footprints shown in Figure 1 reveals two other interesting features. First with the ternary complexes that contain the E54 and the PspFHTH:ADP·AlFx (Figure 1B and C, lane 5b), the o-CuP DNA footprint extends further downstream to about –7 compared to binary E54 complexes (Figure 1B and C, lanes 3a, 4a, 5a and 6a), suggesting that interactions between the E54 and PspFHTH:ADP·AlFx increases interaction with the DNA that is to be melted. DNase I footprints conducted under the same assay conditions used in the o-CuP reactions described above have also shown extended footprints towards the transcription start in E54–PspFHTH:ADP·AlFx ternary promoter complexes . Isomerized binary 54–DNA complexes that are activator and NTP hydrolysis-dependent also produce extended DNase I footprints (19), suggesting that the interaction with the promoter DNA which is melted out in isomerized binary complexes (20) has also occurred in ternary E54–PspFHTH:ADP·AlFx complexes, even though no direct evidence for DNA melting in the ternary complex has been demonstrated (described in the later sections). These results provide further evidence that the bound activator leads either directly or indirectly to changed interactions with promoter DNA downstream of the promoter GC element. Second, in the unbound promoter DNA with the template strand end-labelled from the reactions containing E54 (Figure 1B, lanes 3–6) extra reactivity is seen at –14 and –13 (corresponding to the G and C consensus promoter residues respectively) but not when the non-template strand is labelled (Figure 1C, lanes 3–6). The enrichment of specifically cut template strand sequences in the free DNA suggests that these sequences are attacked by o-CuP when E54 binds promoter DNA to form complexes that do not survive native gel electrophoresis. The extra reactivity at –14 and –13 is unchanged by PspFHTH:ADP·AlFx (Figure 1B, compare lanes 3 and 5), in marked contrast to the –11 reactivity (Figure 1B, compare lanes 3a and 5b), implying that one binary complex is specifically changed in an activator plus ADP·AlFx-dependent way, whereas the unstable binary complex is not.
ADP·AlFx bound to activator changes interactions between promoter DNA and 54
Direct binding of PspFHTH:ADP·AlFx to 54 independent of core RNAP has been demonstrated (25). To determine the extent to which the footprints of E54 bound by the PspFHTH:ADP·AlFx complex were contributed to by 54–PspFHTH:ADP·AlFx, we conducted o-CuP footprints of 54 and its complex with PspFHTH:ADP·AlFx on S.meliloti nifH homoduplex DNA. As shown in Figure 2, compared to the binary 54–DNA complex (lane 3), the ternary 54–DNA complex bound with the ADP·AlFx form of the activator (lane 4a) gives additional protection close to the fork junction region, and some extra reactivity located just downstream of the fork junction region. These patterns of cutting are specific since the promoter DNA isolated from the binary 54–DNA complex isolated from the same binding reaction does not show the pattern of cutting by o-CuP seen in the ternary complex (Figure 2, compare lanes 4b with 4a). Rather the cutting pattern is very close to that seen in reactions that just contain 54 and promoter DNA (Figure 2, compare lanes 4b with 3). Overall results confirmed that the interaction of 54 with PspFHTH:ADP·AlFx results in clear changes in DNA footprints across promoter sequences associated with the fork junction structure found in the E54–promoter DNA complex (Figure 2, compare lanes 4a and 5 respectively). It seems that 54–PspFHTH:ADP·AlFx interactions (Figure 2, lane 4a) mediate the changes in cutting patterns seen with the ternary E54–PspFHTH:ADP·AlFx complex (Figure 1B and 1C, lane 5b).
Figure 2. The o-CuP DNA footprints of ternary 54–PspFHTH:ADP·AlFx promoter complexes. Reactions contain 100 nM S.meliloti nifH homoduplex promoter probe (template strand end-labelled), 2 μM 54 or 250 nM E54, 20 μM PspFHTH and 100 ng/μl salmon sperm DNA. Increased reactivity (filled circles) and decreased reactivity (unfilled circles) to o-CuP in the ternary 54–PspFHTH:ADP·AlFx promoter complex is indicated (lane 4a) as well as the increased reactivity corresponding to the DNA fork junction structure seen in binary E54 complexes (arrowed, lane 5). Lane 4b shows the footprint of the binary 54–DNA complex formed in the same reaction (lane 4a) containing PspFHTH:ADP·AlFx.
Ternary holoenzyme complexes have altered –6, +1 interactions
Nucleotide hydrolysis by the activator is needed for 54 and E54 to melt promoter DNA (19,20,32). To help determine whether ternary 54–PspFHTH:ADP·AlFx promoter complexes might contain melted DNA, we used KMnO4 to probe for single-stranded DNA. Although we demonstrated that the binary E54 promoter complexes were closed complexes the ternary E54–PspFHTH:ADP·AlFx promoter complexes were largely destroyed by KMnO4 treatment. Therefore, we used o-CuP footprinting to see if the PspFHTH:ADP·AlFx-dependent 54 ternary complex could be distinguished from a nucleotide-dependent isomerized binary 54 complex in which promoter melting has been demonstrated (20). A heteroduplex promoter DNA probe based on the E.coli glnHp2 m-12 promoter (33), which creates the artificial fork junction next to the GC required for 54 isomerization, was employed since clear DNA melting is evident at this promoter . The native gel shown in Figure 3A indicates the positions of the binary (54–DNA), nucleotide-dependent isomerized 54–DNA (ss–DNA), and E54–DNA and ternary PspFHTH:ADP·AlFx-dependent (54–DNA and E54–DNA) complexes from which o-CuP-treated heteroduplex promoter DNA (template strand end-labelled) was isolated. In marked contrast to the different o-CuP footprints of binary 54–DNA and E54–DNA complexes seen with the homoduplex probe (Figure 2, lanes 3 and 5 respectively), with heteroduplex DNA (E.coli glnHp2) the binary 54 and E54 footprints are indistinguishable (Figure 3B, compare lanes 3a and 6a). This indicates that the heteroduplex region largely fulfils the role of core RNA polymerase in creating the fork junction. Comparing isomerized 54 (Figure 3B, lane 4b) and ternary 54–PspFHTH:ADP·AlFx (Figure 3B, lane 5b) complexes indicates very close similarities in DNA structure within both complexes. The o-CuP footprints (Figure 3B) also clearly show that in all the binary and ternary complexes the o-CuP reactivity around the region of heteroduplex was greatly diminished compared to the free DNA (Figure 3B, lane 2), suggesting a close interaction between 54 and the DNA fork junction structure in each complex. Comparing E54–DNA (Figure 3B, lanes 7a and 8a) to the ternary ADP·AlFx-dependent complex (Figure 3B, lane 8b) shows that in the E54–PspFHTH:ADP·AlFx complex there is some reduced o-CuP cutting around –6, but markedly at +1. Clearly in 54 and E54 complexes with PspFHTH:ADP·AlFx, the interactions with DNA seem different around the transcription start site (+1) (Figure 3B, compare lanes 4b and 8b), suggesting that this is core RNA polymerase-dependent.
Figure 3. o-CuP DNA footprints of ternary ADP·AlFx-dependent and binary nucleotide-dependent isomerized 54 complexes formed on heteroduplex promoter DNA. (A) Native gel showing the position of binary and ternary complexes from which o-CuP-treated promoter DNA was isolated. (B) o-CuP footprints of DNA isolated from complexes shown in (A). E.coli glnHp2 heteroduplex promoter DNA (see Table 1) was at 100 nM, 54 1 μM, E54 was at 100 nM, PspFHTH at 20 μM and salmon sperm DNA at 100 ng/μl. Isomerized 54 binary complexes were formed in the presence of GTP (4 mM) as the hydrolysable nucleotide. Filled arrowheads indicate the position of the region of heteroduplex that is protected in ternary and binary 54 complexes formed on heteroduplex promoter DNA.
DMS footprints reveal conserved promoter contacts in ternary complexes
Conserved G:C base pairs characterize 54-binding sites, with one set adjacent to the –12 proximal DNA fork junction structure formed in E54 closed promoter complexes, the other set at –24. We conducted DNA footprints using DMS to measure the interactions of 54 and E54 in the absence and presence of ADP·AlFx and PspFHTH to learn if PspFHTH:ADP·AlFx changed the specific sets of 54–DNA interactions involving conserved G:C base pairs at the homoduplex S.meliloti nifH promoter (see Table 1). Methylation protection of the guanines was measured. DMS footprints shown in Figure 4 confirm that the consensus G residues (G-14, G-25 and G-26) of the promoter were protected from DMS attack, as expected in binary 54 and E54 complexes (compare lanes 3b, 3f and 5b with 5f, respectively). Comparison of the 54 and E54 binary complex footprints (Figure 4, lanes 3b and 5b) to those of the 54 and E54 ternary complexes (Figure 4, lanes 4b and 6b respectively) formed with PspFHTH:ADP·AlFx did not show any major changes in the patterns of DMS reactivity. A control with just PspFHTH:ADP·AlFx and promoter DNA (Figure 4, lane 7f) showed that the methylation protection observed was dependent on the presence of 54. To learn if the guanine-specific DNA footprints might usefully distinguish between the ternary 54–DNA complex with bound PspFHTH:ADP·AlFx and the nucleotide-dependent isomerized binary 54 complex, DMS footprints were conducted on S.meliloti nifH DNA with a short region of heteroduplex next to the promoter GC at –12 necessary for efficient isomerization (see Table 1). Comparisons of the DNA footprints from binary 54–DNA and isomerized 54–DNA complexes and the ternary 54–DNA complex showed that conserved guanines were protected from methylation by DMS and that the footprints were indistinguishable (data not shown). Binary 54–RNAP and ternary E54–PspFHTH:ADP·AlFx complexes formed with the heteroduplex promoter DNA also showed similar DMS footprints with conserved guanines being protected from methylation by DMS (data not shown). The DMS footprinting results indicate that key 54–DNA promoter contacts are unchanged by binding activator in the transition state for ATP hydrolysis, suggesting that changes in o-CuP reactivity at –11 (Figure 1B and C) arise when contacts at the major promoter specificity sequences are being maintained.
Figure 4. DMS footprints on homoduplex S.meliloti nifH promoter DNA. Promoter DNA was isolated from unbound (lanes marked with ‘f’) and bound (lanes marked with ‘b’) binary and ternary 54 and E54 complexes following the separation on a native polyacrylamide gel (data not shown). Reactions included 100 nM homoduplex S.meliloti nifH promoter DNA (non-template strand end-labelled), 2 μM 54, 250 nM E54, 20 μM PspFHTH and 100 ng/μl salmon sperm DNA. The position of consensus guanines (G-14, G-25 and G-26) that are protected from methylation in binary and ternary promoter complexes is indicated.
Interactions between activator and DNA
To examine whether there might be a direct activator–DNA interaction within the ternary E54–PspFHTH:ADP·AlFx promoter complexes described above, we conducted o-CuP DNA footprinting reactions using the heteroduplex E.coli glnHp2 DNA template (see Table 1) with increasing amounts of PspFHTH in the absence or presence of ADP·AlFx. The results shown in Figure 5A indicate that increasing the concentration of PspFHTH does lead to a reduced o-CuP reactivity across, and downstream of, the region of heteroduplex implying that there is potential for a DNA contact made by the AAA+ domain of PspFHTH within the ternary complexes. The presence of ADP·AlFx did not appear to have any effect on PspFHTH binding (Figure 5A, compare lanes 4, 6, 8 and 10 with lanes 3, 5, 7 and 9, respectively), and the protein storage buffer did not cause loss of cutting by o-CuP (data not shown). Reduced o-CuP reactivity across, and downstream of the region of heteroduplex on the non-template strand was also observed with PspFHTH and the addition of ADP·AlFx had no effect on PspFHTH binding (data not shown). Forms of PspFHTH that carried different N-terminal tags, or had the His-tag removed also gave a footprint characterized by reduced DNA reactivity downstream of the –12 GC consensus promoter sequence (data not shown) providing evidence that the tags on PspF sequences do not contribute to the pattern of o-CuP cutting seen in the footprinting assays. The reduced DNA reactivity was also seen on the template and non-template strands with a construct comprising amino acids 1–275 of PspF (PspF1–275), the minimum region of PspF corresponding to the AAA+ domain (Figure 5B, left-hand panel and data not shown). Therefore, the o-CuP footprint is not due to sequences that link the AAA+ domain of PspF to its DNA-binding domain. Similar results were obtained when we replaced the heteroduplex E.coli glnHp2 DNA with the S.meliloti nifH heteroduplex (see Table 1) promoter probe, suggesting that PspF1–275 may contact 54-dependent promoters in general (data not shown). We repeated the o-CuP footprinting reactions shown in Figure 5A but replaced the heteroduplex E.coli glnHp2 promoter probe with homoduplex E.coli glnHp2 promoter DNA (see Table 1). The homoduplex E.coli glnHp2 probe alone displayed strong o-CuP reactivity between about –6 and –9, indicating that a particular minor groove geometry there expected from its A:T richness (Figure 5C, compare lanes 1 and 2). An interaction of PspFHTH with the E.coli glnHp2 homoduplex DNA was also evident as judged by reduced o-CuP reactivity downstream of the –12 GC (see Figure 5C), indicating that the region of heteroduplex was not required for the binding of PspFHTH.
Figure 5. Activator interacts with heteroduplex and homoduplex E.coli glnHp2 promoter DNA. (A) o-CuP DNA footprints resulting from the addition of increasing amounts of PspFHTH (2, 5, 10 and 20 μM) to 100 nM heteroduplex E.coli glnHp2 promoter DNA (see Table 1) in the absence or presence of ADP·AlFx. (B) o-CuP footprints with pspF1–275 at 2, 5, 10 and 20 μM. (C) o-CuP footprints as in (A), except that homoduplex E.coli GlnHp2 promoter probe (see Table 1) was used. Salmon sperm DNA was not included in the binding reactions. The region of heteroduplex is indicated by filled arrowheads, and the promoter region showing reduced o-CuP reactivity in the presence of activator is indicated by a solid line. Lane ‘m’, size marker generated by end-labelling oligonucleotides.
To examine whether the DNA footprints due to the presence of PspF alone were specific to certain PspF residues, increasing amounts (5–20 μM) of mutant PspF1–275 (the minimal AAA + domain of PspF) proteins (25,28,31) were added to the heteroduplex E.coli glnHp2 promoter probe and the DNA examined by o-CuP footprinting. The results (data not shown) show that the single amino acid substitution mutant PspF1–275 proteins tested (K42A, T86A, T86S, D107A, R122A, T148A, N149A, R162A, R168A, R227A, K230A and N231A) with defects in the GAFTGA, Walker A, Walker B motifs and so-called Sensor regions (22) are not significantly defective (at most a 2-fold reduction in binding for the D107A mutant) in promoter interaction. Similar to PspF1–275, the mutant PspF1–275 proteins showed similar protection downstream of the –12 GC consensus promoter sequence from o-CuP cutting, suggesting that the substituted amino acid residues do not contribute to the DNA-binding activity of the activator.
In order to determine whether the o-CuP footprint, we observed with the AAA+ domain of PspF was evident with another protein we purified the AAA+ domain from the 54-dependent activator, C4-dicarboxylic acid transport protein D (DctD) from S.meliloti (see Materials and Methods). The results using the heteroduplex E.coli glnHp2 promoter probe with the template strand end-labeled (Figure 5B, right-hand panel) or non-template strand end-labeled (data not shown) clearly show that the DctD AAA+ domain gives a similar o-CuP footprint to that of PspF1–275 (Figure 5B, left-hand panel and data not shown) and protects the DNA from o-CuP cutting, across and downstream from the region of heteroduplex. The presence of ADP·AlFx had no effect upon DctD binding (data not shown).
When the binding reactions between activator and DNA in the absence or presence of ADP·AlFx (see Materials and Methods) were run on native polyacrylamide gels no strong clear band shifted binary activator–DNA complexes were seen (see Figure 1A, lane 10 and Figure 3A, lane 9 and data not shown), suggesting that the DNA binding that gave rise to the o-CuP footprints was very unstable.
A composite 54, PspFHTH DNA footprint
Consistent with the binding of PspFHTH to promoter DNA being weak, titration of salmon sperm DNA into reactions containing promoter probe and PspFHTH resulted in a gradual loss of protection, with no protection being seen above 100 ng/μl salmon sperm DNA (data not shown). Recall that salmon sperm DNA was included in the earlier o-CuP and DMS footprints (Figures 1–4), and can explain why the control reactions in those experiments (i.e. probe plus PspFHTH) do not reveal a direct interaction of PspFHTH with DNA. As shown in Figure 6, inclusion of 54 in binding reactions with PspFHTH and the E.coli glnHp2 promoter probe (template strand end-labelled) in the absence of salmon sperm DNA (lane 6) resulted in an o-CuP footprint that was apparently a composite of the 54 footprint (lane 3) and that due to PspFHTH alone (lane 8). Under conditions where the 54 protein stably bound to the activator through the inclusion of ADP·AlFx a composite o-CuP footprint was again evident (Figure 6, lane 7). A similar composite footprint was also observed in the nucleotide-dependent isomerized binary 54–DNA complex (Figure 6, lane 4). Although composite footprints are observed independent of the nucleotide (lanes 4, 6 and 7), only PspFHTH in the presence of ADP·AlFx forms stable ternary complexes, suggesting different structural states that are nucleotide-dependent. Addition of salmon sperm DNA diminished the protection associated with PspFHTH alone, but the protection due to 54 persisted (see Figure 3B), suggesting that the interaction between the fork junction DNA and the activator is weak, and if occurring in ternary promoter complexes, would have to be stabilized by interactions dependent upon 54 and E54.
Figure 6. ‘Composite’ o-CuP DNA footprints of ternary ADP·AlFx-dependent and binary nucleotide-dependent isomerized 54 complexes formed on heteroduplex E.coli glnHp2 promoter DNA in the absence of salmon sperm DNA. Reactions were essentially as in Figure 3 but salmon sperm DNA was not included in the binding reactions. Filled arrowheads indicate the position of the region of heteroduplex that is protected in ternary and binary 54 complexes formed on heteroduplex promoter DNA.
DNA cross-linking to PspF1–275 within a ternary complex
To further investigate possible activator interactions with promoter DNA, we used photoreactive DNA probes (29). The promoter probes were based on the heteroduplex S.meliloti nifH promoter (see Table 1) with a photoactivatable cross link (azidophenacyl bromide) tethered to the DNA via a uniquely positioned phosphorothioate residue either upstream or downstream of the region of heteroduplex (i.e. between bases, –15/–14, –14/–13, –13/–12, –12/–11, –11/–10 and –7/–6) and so including positions that are protected in o-CuP DNA footprints described above, with the phosphorothiolated strand 32P-end-labelled.
Initially, we incubated 54 alone and 54 in the presence of PspF1–275:ADP·AlFx using a DNA probe with the phosphorothioate residue located between bases –7 and –6 on the template strand. The binding reactions were then irradiated prior to loading onto a native polyacrylamide gel (Figure 7, left-hand panel). Binary 54–DNA and ternary 54–PspF1–275:ADP·AlFx complexes clearly formed with the cross-linking probe (Figure 7, labelled ‘b’ in lane 2, and ‘c’ in lane 3, respectively). Control reactions with either PspF1–275 alone (Figure 7, lane 4) or PspF1–275 in the presence of ADP·AlFx (data not shown) revealed no gel-shifted complexes following UV cross-linking. The binary and ternary complexes and unbound DNA resolved on the native gel (Figure 7, lanes 1–4, labelled a–d) were eluted into 1x Laemmli sample buffer (see Materials and Methods) and then further analysed by SDS–PAGE (Figure 7, right-hand panel, lanes a–d). In marked contrast to the binary 54–DNA complex, analysis of the isolated ternary 54–PspF1–275:ADP·AlFx complex on SDS–PAGE showed the presence of an additional protein–DNA cross-linked complex (Figure 7, compare lanes b and c). This band is PspF1–275 cross-linked to the DNA probe (PspF1–275–DNA) as established by using different molecular weight forms of PspF (Burrows et al., manuscript in preparation). The PspF1–275–DNA cross-linked complex (Figure 7, lane c) was only observed when both 54 and ADP·AlFx were present with PspF1–275. PspF1–275 cross-linked to DNA in the presence of RNAP holoenzyme, demonstrating PspF1–275 proximity to DNA in a complex competent for transcriptional activation (30). Also, similar results were obtained with all the cross-linking DNA probes tested (outlined above) with the phosphorothioate residue located on either the template or non-template strand (data not shown). The cross-linking results suggest that within the ternary 54–PspF1–275:ADP·AlFx complex PspF1–275 is proximal to regions of the DNA either side of the region of heteroduplex.
Figure 7. DNA-crosslinking to PspF1–275 is dependent on 54 and ADP·AlFx. Reactions contain a modified heteroduplex S.meliloti nifH promoter probe (100 nM) based on the heteroduplex probe shown in Table 1, but with a photoactivatable cross link positioned between bases –7 and –6 on the template strand. Where indicated PspF1–275 (20 μM), 54 (1 μM) and ADP·AlFx (0.2 mM) are included in the DNA-binding assay. Reactions were then irradiated for 5 min and loaded onto a 4.5% native polyacrylamide gel (left-hand panel). The binary 54–DNA (b, lane 2) and ternary 54–PspF1–275:ADP·AlFx (c, lane 3) complexes, the free DNA (a, lane 1) and the free DNA from the reaction containing PspF1–275 (d, lane 4) were further analysed by SDS–PAGE (right-hand panel).
DNA stimulates the activity of the PspF AAA+ domain, suggesting a functional role for DNA
To help determine a functional relevance for the activator interaction with DNA, we measured the ATPase activity of PspF1–275 in the absence or presence of DNA. In the presence of either homoduplex or heteroduplex E.coli glnHp2 promoter DNA (see Table 1), PspF1–275 repeatedly showed a modest but clear increase in ATPase activity compared to PspF1–275 alone (Figure 8). The increase in activity was not due to ATPase activity of the DNA alone (Figure 8 and data not shown). Next, to determine if the stimulation is promoter sequence specific, we carried out the same experiments in the presence of salmon sperm DNA. The same level of ATPase stimulation was observed with salmon sperm DNA, demonstrating that the stimulation of PspF1–275 is DNA-dependent but not obviously DNA sequence-dependent. This is in agreement with our DNA foot-printing experiments where salmon sperm DNA quenches DNA protection of the promoter DNA. These observations suggest that relatively weak and unspecific interactions between the AAA+ domain of PspF (PspF1–275) and the promoter DNA in the ternary E54–DNA complex may play a functional role by contributing to productive interactions within a DNA-dependent architecture of the initiation complex. Although PspF1–275–DNA interactions are unspecific, 54–PspF1–275 interactions could direct PspF proximal to the promoter sequence. We considered that DNA could increase the ATPase activity of PspF1–275 by promoting higher order oligomer formation of PspF1–275 or by increasing the activity of the catalytic site of PspF1–275 subunits. Evidence for an effect upon higher order oligomer formation comes from the finding that stimulation by DNA is dependent upon PspF concentration in the range where PspF1–275 oligomerizes (31) and is independent of ATP concentration (data not shown).
DISCUSSION
The activators of E54 are mechanochemical ATPases that use ATP hydrolysis to promote DNA strand opening and stimulate the associated changes in 54 structure needed for open complex formation (6,11,19,22,25). The ways in which nucleotide binding and its hydrolysis used for open complex formation are not well understood, but the effects of ADP·AlFx presented here begin to address this issue. Several lines of evidence, notably the use of DNA fork junction promoter probes and related heteroduplex probes, have shown that the E54 makes use of the double-strand–single-strand junction to limit DNA opening (8,11,12,19,20,23,26). Changed E54 interactions at the fork junction appear to be required to allow open complexes to form (19,34) and our new data strongly support and extend this view. Variations in promoter sequences suggest that a range of natural DNA fork junctions will exist. The functional state of the activator required for stable binding to the E54 was created by interaction with ADP·AlFx, an analogue of ATP at the transition state for hydrolysis. The N-terminal Region I of 54 is required for creating the fork junction structure and activator binds directly to Region I, suggesting that activator changes interactions between the fork junction structure and 54 in a nucleotide-dependent manner (25). A comparison (Figure 1B) of bound and unbound DNA exposed to E54 showed that the effects of PspFHTH:ADP·AlFx were restricted to complexes resolved in native gels, and that the effects of PspFHTH:ADP·AlFx upon unstable E54 promoter DNA complexes were not evident. Consideration of previous o-CuP footprinting (14) carried out in solution and without resolving complexes indicate that these footprints in fact arise from a mixture of stable complexes (o-CuP signal in the –11 region) and unstable complexes (o-CuP signal in the –14 region).
The changed DNA interactions at the fork junction structure (which may be entirely indirect and through 54, or may involve a DNA contact by the AAA+ domain of PspF) suggest that a part of the DNA opening pathway involves an activator interaction with promoter DNA, potentially either side of the fork junction structure. The PspF AAA+ domain showed clear protection around the region of heteroduplex present in the DNA probes used in this work. A number of conserved residues in PspF appeared not to contribute to the DNA binding shown by the minimal AAA+ domain, implying that interactions of PspF1–275 with promoter DNA will not greatly change the activities of PspF associated with these conserved amino acids. Some other AAA+ proteins notably those related to RuvB that remodel DNA and nucleoprotein complexes are thought to make direct interactions with the DNA component of their targets. In these cases, a DNA-dependent stimulation of the ATPase function has been reported, and we have also observed this (see Figure 8) for the minimal AAA+ domain of PspF (PspF1–275) and for PspFHTH (data not shown) in the presence of promoter DNA and non-specific DNA. This supports the fact that weak but direct PspF1–275 contacts to the DNA occur and are significant for the nucleotide-dependent formation of open promoter complexes. Direct evidence for AAA+ domain–DNA interactions comes from the recently published structures from replication factor C of the clamp loader complex, involved in DNA replication (38,39). In our current model, specific DNA binding is mediated by the PspF HTH motif to its upstream enhancer site. However, in the ternary E54–PspF1–275:ADP·AlFx–DNA complex, PspF1–275 (the AAA+ domain) is proximal to the promoter DNA where it could contribute to the architecture of the ternary complex by interacting with DNA sequences across and downstream of the –12 GC. In the context of the full-length PspF protein, the downstream promoter interactions by the AAA+ domain may be enhanced by the specific DNA-binding activity of the HTH motif. upstream. Qualitative aspects of the footprint displayed by PspF1–275 (Figure 5B) were independent of any functional state we examined, including the ADP·AlFx bound state, the ADP bound state and 54 Region I bound state (data not shown).
A related architecture between RuvB (35) and PspF is predicted (30,36) and some relatedness in where these proteins interact with their targets is expected. Interactions between the Region I of 54, the fork junction DNA and PspF might therefore be direct, and help to modify the ATPase domain structure and activity to enable remodelling of the closed promoter complex. RuvA could interact with RuvB as 54 interacts with its activators. RuvB and PspF could contact their respective DNA targets.
A common core RNA polymerase enzyme is used by the enhancer-independent class of sigma factors. It would appear that the special features of 54-dependent transcription relate closely to the activator targeting an unusually conformationally stable fork junction containing complex by making contact with the DNA and protein components of the closed complex. Some of the 54–DNA interactions that activator changes seem to have a modest energetic cost, as evidenced by the action of the ATPS bound activator (26). Others, notably the DNA opening per se seem to correlate to a full ATP hydrolysis event and appear to have a higher energetic cost. The sensing of the gamma phosphate of the ATP is implicated as critically changing the functionality of the activator (22), a common theme for AAA+ proteins where nucleotide binding and hydrolysis control the binding interactions needed for substrate remodelling. For activation of promoters by the catabolite activating protein CAP, remodelling of promoter complexes occurs, but uses an activator–RNA polymerase contact upstream of the main promoter sequences (37). The use of DNA looping and enhancers appears to enable the 54 activators to make a contact just downstream of the main promoter elements, near or at the place where the fork junction exists in closed complexes. It now appears that core promoter DNA is contacted by the AAA+ domain of the activator, in addition to the promoter contacts made by 54.
ACKNOWLEDGEMENTS
We thank P.Bordes for PspF proteins, P.Ray for 54 protein and T. Hoover for plasmid pHX187 encoding the S.meliloti DctD AAA+ domain. The DNA probes used in cross-linking experiments were designed and prepared by S. Wigneshweraraj and P.Burrows, whose help is gratefully acknowledged. We also thank S.Wigneshweraraj and G.Jovanovic for comments on the manuscript. Work was supported by a Wellcome Trust project grant to M.B.
REFERENCES
Ebright,R.H. ( (2000) ) RNA polymerase: structural similarities between bacterial RNA polymerase and eukaryotic RNA polymerase II. J. Mol. Biol., , 304, , 687–698.
Cramer,P. ( (2002) ) Multisubunit RNA polymerases. Curr. Opin. Struct. Biol., , 12, , 89–97.
Gross,C.A., Chan,C., Dombroski,A., Gruber,T., Sharp,M., Tupy,J. and Young,B. ( (1998) ) The functional and regulatory roles of sigma factors in transcription. Cold Spring Harb. Symp. Quant. Biol., , 63, , 141–155.
Burgess,R.R. and Anthony,L. ( (2001) ) How sigma docks to RNA polymerase and what sigma does. Curr. Opin. Microbiol., , 2, , 126–131.
Studholme,D.J. and Buck,M. ( (2000) ) The alternative sigma factor sigma(28) of the extreme thermophile Aquifex aeolicus restores motility to an Escherichia coli fliA mutant. FEMS Microbiol. Lett., , 191, , 103–107.
Buck,M., Gallegos,M.-T., Studholme,D.J., Guo,Y. and Gralla,J.D. ( (2000) ) The bacterial enhancer-dependent 54 (N) transcription factor. J. Bacteriol., , 15, , 4129–4136.
Xu,H. and Hoover,T.R. ( (2001) ) Transcriptional regulation at a distance in bacteria. Curr. Opin. Microbiol., , 4, , 138–144.
Guo,Y., Wang,L. and Gralla,J.D. ( (1999) ) A fork junction DNA–protein switch that controls promoter melting by the bacterial enhancer-dependent sigma factor. EMBO J., , 18, , 3736–3745.
Guo,Y. and Gralla,J.D. ( (1998) ) Promoter opening via a DNA fork junction binding activity. Proc. Natl Acad. Sci. USA, , 95, , 11655–11660.
Chaney,M. and Buck,M. ( (1999) ) The 54 DNA-binding domain includes a determinant of enhancer responsiveness. Mol. Microbiol., , 33, , 1200–1209.
Guo,Y., Lew,C.M. and Gralla,J.D. ( (2000) ) Promoter opening by sigma(54) and sigma(70) RNA polymerases: sigma factor-directed alterations in the mechanism and tightness of control. Genes Dev., , 14, , 2242–2255.
Wedel,A. and Kustu,S. ( (1995) ) The bacterial enhancer-binding protein NTRC is a molecular machine: ATP hydrolysis is coupled to transcriptional activation. Genes Dev., , 9, , 2042–2052.
Cannon,W., Gallegos,M.-T., Casaz,P. and Buck,M. ( (1999) ) Amino-terminal sequences of N (54) inhibit RNA polymerase isomerization. Genes Dev., , 13, , 357–370.
Morris,L., Cannon,W., Claverie-Martin,F., Austin,S. and Buck,M. ( (1994) ) DNA distortion and nucleation of local DNA unwinding within 54 (N) holoenzyme closed promoter complexes. J. Biol. Chem., , 269, , 11563–11571.
Wigneshweraraj,S.R., Konstantin,K., Konstantin,S. and Buck,M. ( (2003) ) Multiple roles of the RNA polymerase ? subunit flap domain in54-dependent transcription. J. Biol. Chem., , 278, , 3455–3465.
Murakami,K.S., Masuda,S. and Darst,S.A. ( (2002) ) Structural basis of transcription initiation: RNA polymerase holoenzyme at 4 ? resolution. Science, , 296, , 1280–1284.
Murakami,K.S., Masuda,S., Campbell,E.A., Muzzin,O. and Darst,S.A. ( (2002) ) Structural basis of transcription initiation: an RNA polymerase holoenzyme–DNA complex. Science, , 296, , 1285–1290.
Tsujikawa,L., Tsodikov,O.V. and deHaseth,P.L. ( (2002) ) Interaction of RNA polymerase with forked DNA: evidence for two kinetically significant intermediates on the pathway to the final complex. Proc. Natl Acad. Sci. USA, , 99, , 3493–3498.
Cannon,W.V., Gallegos,M.-T. and Buck,M. ( (2000) ) Isomerization of a binary sigma–promoter DNA complex by transcription activators. Nature Struct. Biol., , 7, , 594–601.
Cannon,W., Gallaegos,M.-T. and Buck,M. ( (2001) ) DNA melting within a binary sigma(54)–promoter DNA complex. J. Biol. Chem., , 276, , 386–394.
Rombel,I., North,A., Hwang,I., Wyman,C. and Kustu,S. ( (1998) ) The bacterial enhancer-binding protein NtrC as a molecular machine. Cold Spring Harb. Symp. Quant. Biol., , 63, , 157–66.
Zhang,X., Chaney,M., Wigneshweraraj,S.R., Schumacher,J., Bordes,P., Cannon,W. and Buck,M. ( (2002) ) Mechanochemical ATPases and transcriptional activation. Mol. Microbiol., , 45, , 895–903.
Cannon,W., Wigneshweraraj,S.R. and Buck,M. ( (2002) ) Interactions of regulated and deregulated forms of the sigma54 holoenzyme with heteroduplex promoter DNA. Nucleic Acids Res., , 30, , 886–893.
Wang,L., Guo,Y. and Gralla,J.D. ( (1999) ) Regulation of sigma54-dependent transcription by core promoter sequences: role of –12 region nucleotides. J. Bacteriol., , 181, , 7758–7765.
Chaney,M., Grande,R., Wigneshweraraj,S.R., Cannon,W., Casaz,P., Gallegos,M.T., Schumacher,J., Jones,S., Elderkin,S., Dago,A.E., Morett,E. and Buck,M. ( (2001) ) Binding of transcriptional activators to sigma54 in the presence of the transition state analog ADP–aluminium chloride: insights into activator mechanochemical action. Genes Dev., , 15, , 2282–2294.
Cannon,W., Bordes,P., Wigneshweraraj,S.R. and Buck,M. ( (2003) ) Nucleotide-dependent triggering of RNA polymerase–DNA interactions by an AAA regulator of transcription. J. Biol. Chem., , 278, , 19815–19825.
Jovanovic,G., Rakonjac,J. and Model,P. ( (1999) ) In vivo and in vitro activities of the Escherichia coli sigma54 transcription activator, PspF, and its DNA-binding mutant, PspFDeltaHTH. J. Mol. Biol., , 285, , 469–483.
Bordes,P., Wigneshweraraj,S.R., Schumacher,J., Zhang,X., Chaney,M. and Buck,M. ( (2003) ) The ATP hydrolyzing transcription activator phage shock protein F of Escherichia coli: identifying a surface that binds 54. Proc. Natl Acad. Sci. USA, , 100, , 2278–2283.
Mayer,A.N. and Barany,F. ( (1995) ) Photoaffinity cross-linking of TaqI restriction endonuclease using an aryl azide linked to the phosphate backbone. Gene, , 153, , 1–8.
Burrows,P.C., Serinov,K., Buck,M. and Wigneshweraraj,S.R. ( (2004) ) Re-organisation of an RNA polymerase-promoter DNA complex for DNA melting. EMBO J., , in press.
Schumacher,J., Zhang,X., Jones,S., Bordes,P. and Buck,M. ( (2004) ) ATP-dependent transcriptional activation by bacterial PspF AAA+ protein. J. Mol. Biol., , 338, , 863–875.
Popham,D., Szesto,D., Keener,J. and Kustu,S. ( (1989) ) Function of a bacterial activator protein that binds to transcriptional enhancers. Science, , 243, , 629–635.
Claverie-Martin,F. and Magasanik,B. ( (1992) ) Positive and negative effects of DNA bending on activation of transcription from a distant site. J. Mol. Biol., , 227, , 996–1008.
Wigneshweraraj,S.R., Chaney,M.K., Ishihama,A. and Buck,M. ( (2001) ) Regulatory sequences in sigma 54 localise near the start of DNA melting. J. Mol. Biol., , 306, , 681–701.
Han,Y.W., Iwasaki,H., Miyata,T., Mayanagi,K., Yamada,K., Morikawa,K. and Shinagawa,H. ( (2001) ) A unique beta-hairpin protruding from AAA+ ATPase domain of RuvB motor protein is involved in the interaction with RuvA DNA recognition protein for branch migration of Holliday junctions. J. Biol. Chem., , 276, , 35024–35028.
Lee,S.-Y., De La Torre,A., Yan,D., Kustu,S., Nixon,B.T. and Wemmer,D.E. ( (2003) ) Regulation of the transcriptional activator NtrC1: structural studies of the regulatory and AAA+ ATPase domains. Genes Dev., , 17, , 2552–2563.
Chen,H., Tang,H. and Ebright,R.H. ( (2003) ) Functional interaction between RNA polymerase alpha subunit C-terminal domain and sigma70 in UP-element-and activator-dependent transcription. Mol. Cell, , 11, , 1621–1633.
Miyata,T., Oyama,T., Mayanagi,K., Ishino,S., Ishino,Y. and Morikawa,K. ( (2004) ) The clamp-loading complex for processive DNA replication. Nature Struct. Biol., , 11, , 632–636.
Bowman,G.D., O'Donnell,M. and Kuriyan,J. ( (2004) ) Structural analysis of a eukaryotic sliding DNA clamp–clamp loader complex. Nature, , 429, , 724–730.(W. V. Cannon, J. Schumacher and M. Buck*)
* To whom correspondence should be addressed. Tel: +44 207 594 5442; Fax: +44 207 594 5419; Email: m.buck@imperial.ac.uk
ABSTRACT
Enhancer-dependent transcriptional activators that act upon the 54 bacterial RNA polymerase holoenzyme belong to the extensive AAA+ superfamily of mechanochemical ATPases. Formation and collapse of the transition state for ATP hydrolysis engenders direct interactions between AAA+ activators and the 54 factor, required for RNA polymerase isomerization. A DNA fork junction structure present within closed complexes serves as a nucleation point for the DNA melting seen in open promoter complexes and restricts spontaneous activator-independent RNA polymerase isomerization. We now provide physical evidence showing that the ADP·AlFx bound form of the AAA+ domain of the transcriptional activator protein PspF changes interactions between 54-RNA polymerase and a DNA fork junction structure present in the closed promoter complex. The results suggest that one functional state of the nucleotide-bound activator serves to alter DNA binding by 54 and 54-RNA polymerase and appears to drive events that precede DNA opening. Clear evidence for a DNA-interacting activity in the AAA+ domain of PspF was obtained, suggesting that PspF may make a direct contact to the DNA component of a basal promoter complex to promote changes in 54-RNA polymerase–DNA interactions that favour open complex formation. We also provide evidence for two distinct closed promoter complexes with differing stabilities.
INTRODUCTION
The interactions that RNA polymerases make with transcriptional control proteins and how these proteins interface with signalling pathways is central to achieving appropriate patterns of gene expression. The multisubunit DNA-dependent RNA polymerases (RNAPs) of bacteria and eukaryotes are closely related in structure, having a common line of evolutionary descent (1,2). In bacteria, core RNAP (2??'; E) associates with one of several sigma () factors to form functional holoenzymes (2??'; E) capable of recognizing specific promoter DNA targets (3,4).
Many bacteria contain the enhancer-dependent 54 factor (5–7), and the RNA polymerase holoenzyme containing 54 (E54) has the distinctive property of remaining bound to a promoter in a closed complex that very rarely spontaneously isomerizes to an open complex (6,7). A network of interactions among 54, core RNAP and the promoter DNA are required to stabilize the closed promoter complex conformation (8–10) and hence limit isomerization and formation of open complexes prior to activation (11–13). A DNA fork junction persists within the E54 closed complex, and interactions of 54 with this DNA structure are thought to contribute to maintaining the closed promoter complex (8,11,14,15). Weak binding of holoenzyme containing the 70 (E70) type factor to a fork junction structure is believed to be related to the ability of the E70 to transcribe without the strict requirement of an activator protein (11). Current RNA polymerase promoter complex structures (2,16,17) suggest that RNAP interactions at the fork junction will contribute to delivering the promoter DNA to be melted into the DNA-binding cleft of the core RNAP enzyme.
The use of fork junction containing promoter DNA fragments has greatly enabled structural (16,17) and functional (8,9,18) studies on transcription initiation. Direct recognition of the DNA fork junction structure formed around position –12 with respect to the transcription start site (+1) by 70 and 54 occurs as part of the process of promoter melting. For both classes of factor, protein and DNA isomerization events occur that lead to the formation of open complexes (11,18,19). Isomerization of the basal 54–DNA complex requires ATP hydrolysis by mechanochemical enhancer-binding transcription activator proteins, which changes its interaction with the fork junction DNA and leads to local DNA denaturation (19,20). Activators of E54 belong to the large AAA+ (ATPase Associated Activities) protein family that use ATP binding and hydrolysis to remodel their substrates (19,21,22). Certain DNA fork junction sequences diminish DNA binding by 54 and play a role in restricting isomerization of the closed promoter complex (23,24). The N-terminal Region I of 54 has been shown to be important for tight binding to DNA fork junction structures (8,11,13). The activators bind to the Region I of 54 and use ATP hydrolysis to change fork junction structure interactions made by 54 (19,25). The stable binding of 54 to its activators can be detected in the presence of ADP·AlFx, a transition state analogue of ATP thought to change the functional state of the activator leading to its stable engagement with 54 (25).
Nucleotide-dependent putative intermediates in the 54 transcription activation pathway have recently been identified (26). We now show that ADP·AlFx-dependent binding of the ATP hydrolysing AAA+ domain of the Escherichia coli phage shock protein PspF to E54 changes interactions between the E54 and the fork junction DNA structure present within the closed promoter complex. Interactions between the fork junction DNA and E54 are clearly altered when activator adopts a functional state induced by its interaction with ATP at the point of hydrolysis. A DNA-binding activity within the AAA+ domain of PspF was also evident, indicating that the activator not only makes a direct contact to the 54 protein but may also make contact with a DNA component of a promoter complex. This view is substantiated by the observation that DNA alone stimulates the ATPase activity of the PspF AAA+ domain, where the PspF sequence-specific DNA-binding domain has been deleted. Therefore, 54-mediated interactions between the PspF AAA+ domain and promoter DNA are likely to be of functional importance, and distinct from the tethering of PspF via a DNA-specific binding domain to upstream promoter sequences. Overall the results provide evidence that activation of E54 proceeds sequentially through activator using nucleotide-dependent interactions to remodel a fork junction–E54 complex that then leads to stable DNA opening.
MATERIALS AND METHODS
Proteins
The Klebsiella pneumoniae 54 protein (amino acids 1–477) was prepared as described (23). C-terminal deleted forms of activator PspF lacking a functional DNA-binding domain, PspFHTH (amino acids 1–292) and PspF1–275 (the minimal PspF AAA domain) obtained by overproducing an N-terminal histidine-tagged fragment were used in activation assays (27,28).
The N-terminal histidine-tagged DctD AAA+ domain was overexpressed in E.coli strain ‘Rosetta’ (Novagen) from plasmid pHX187 (a gift from T. Hoover). Freshly transformed cells containing pHX187 were added directly from the agar plate into one litre of ‘Hyper’ media (Athena Enzyme Systems) and grown by shaking at 37°C with good aeration until an OD of 0.7 before induction with 1 mM isopropyl-?-D-galactopyranoside for 3 h. Cells were harvested by centrifugation, resuspended in Buffer A (20 mM Tris, pH 8.0, 65 mM potassium thiocyanate and 5% glycerol) containing ‘Complete’ protease inhibitor cocktail (Roche) and lysed in a cell disrupter at 25 KPSI. The cell lysate was clarified by centrifugation at 15 000 r.p.m. for 30 min at 4°C. The resulting supernatant was applied to a HiTrap 5 ml Nickel chelating column and the His-tagged DctD AAA+ domain eluted from the column by running a 40 ml gradient between 5 and 100% Buffer B (Buffer A containing 1 M imidazole). Column fractions containing the His-tagged DctD AAA+ domain were dialysed overnight into 10 mM Tris, pH 8.0, 50% glycerol, 0.1 mM EDTA, 1 mM DTT and 50 mM NaCl at 4°C, protein concentration measured and protein stored at –80°C.
E.coli core RNA polymerase (E) was purchased from Epicentre Technologies Corp.
DNA duplex formation
Wild-type homoduplex or mismatched Sinorhizobium meliloti nifH and E.coli glnHp2 promoter DNA heteroduplexes were prepared by annealing 88mer oligonucleotides . Pairs of oligonucleotides with one strand 5'-32P-end-labelled and with the unlabelled strand at a 2-fold molar excess were heated at 95°C for 3 min in 10 mM Tris–HCl (pH 8.0), 10 mM MgCl2 and then chilled rapidly in iced H2O for 5 min to allow duplex formation.
Table 1. S.meliloti nifH and E.coli glnHp2 promoter duplexes
Native gel mobility shift assays
Gel mobility shift assays were employed to detect 54 and E54 binary and ternary complexes formed with radioactively labelled S.meliloti nifH and E.coli glnHp2 duplex promoter fragments. 54 (1 μM) or E54 were incubated with 32P-end-labelled duplex DNA (100 nM) at 30°C for 5 min in STA buffer . When required activator (2–20 μM) was added for a further 10 min in the absence or presence (2–4 mM) of a nucleoside triphosphate (GTP), or ADP·AlFx formed in situ by the addition of 0.2 mM AlCl3 to a mixture containing 0.2 mM ADP and 5.0 mM sodium fluoride (25). Unless stated otherwise, non-specific competitor DNA (salmon sperm DNA, 100 ng/μl) was added to binding reactions (see figure legends). A glycerol bromophenol blue loading dye (final concentration 10% glycerol) was added to the samples which were then loaded onto 4.5% native polyacrylamide gels and run at 60 V for 80 min at room temperature in 25 mM Tris and 200 mM glycine buffer (pH 8.6). Unbound DNA and protein–DNA complexes were detected by autoradiography.
DNA footprinting assays
Binding reactions (10 μl) containing 100 nM of 32P-end-labelled duplex DNA were conducted as described above. Footprinting reagents were then added; reactions terminated as indicated below; and bound and unbound DNAs were separated on native polyacrylamide gels. DNA was then excised and eluted overnight. For ortho-copper phenanthroline (o-CuP) footprints 0.5 μl of 0.45 mM CuSO4/2.0 mM 1,10-phenanthroline monohydrate and 0.5 μl of 116 mM 2-mercaptopropionic acid (to start the reaction) were added to the binding assay. After 3 min, the reaction was terminated by quenching with 1.0 μl of 28 mM 2,9-dimethyl-1,10-phenanthroline. For dimethylsulphate (DMS) footprints, 1.0 μl of 5% (v/v) DMS was added to the binding reaction for 2 min after which time the reaction was terminated by adding 50 mM 2-mercaptoethanol. Gel-isolated DNA was eluted into H2O overnight at 37°C. DMS-treated DNA was then cleaved with piperidine (10% v/v) at 90°C for 20 min and the DNA pellet washed with H2O. Recovery of isolated DNA was determined by dry Cerenkov counting, and equal numbers of counts were loaded onto denaturing 10% polyacrylamide gels.
DNA cross-linking assays
Promoter probes were based on the heteroduplex S.meliloti nifH promoter (see Table 1) with a photoactivatable cross link (azidophenacyl bromide) tethered to the DNA via a uniquely positioned phosphorothioate residue (29) either upstream or downstream of the region of heteroduplex (i.e. between bases, –15/–14, –14/–13, –13/–12, –12/–11, –11/–10 and –7/–6) with the phosphorothiolated strand 32P-end-labelled (30). Binding reactions (10 μl) containing 100 nM of 32P-end-labelled cross-linking DNA probe were conducted as described above. Reactions were then irradiated (wavelength of 365 nm) for 5 min (UV Stratalinker 1800; Stratagene) prior to loading onto 4.5% native polyacrylamide gels. Stable complexes and unbound DNA from the native gel were eluted into 1x Laemmli sample buffer (Sigma) for 1 h at 37°C and then further analysed by 12.5% SDS–PAGE.
ATPase activity assays
ATPase activity assays of PspF1–275 in the absence and presence of DNA were carried out as described in (31). Reactions containing 0.8 μM PspF1–275 or 0.8 μM PspF1–275 and 10 μM E.coli glnHp2 probes (see Table 1) or 0.45 μg/μl salmon sperm DNA were started by adding ATP to a final concentration of 4 mM containing 0.06 μCi/μl . Samples were incubated at 30°C and time-points were taken as indicated (see Figure 8). Reactions were stopped by adding 5 vol of 2 M formic acid. Released ADP was separated from ATP by thin layer chromatography and amounts of radiolabelled ADP and ATP measured by phosphorimaging (Fuji Bas-1500, Tina 2.10g software).
Figure 8. Time-course showing ATP hydrolysis activity of PspF1–275 in the absence or presence of DNA. Homoduplex or heteroduplex E.coli GlnHp2 promoter probes, (see Table 1; diamonds or open circles, respectively), or salmon sperm DNA (triangles) were incubated with PspF1–275 prior to starting reactions by adding ATP (see Materials and Methods) and compared to PspF1–275 in the absence of DNA (squares). Homoduplex E.coli GlnHp2 promoter probe alone (stars) did not show ATP hydrolysis activity.
RESULTS
Activator bound to ADP·AlFx changes interactions between the DNA fork junction and E54
We used the minor groove-specific DNA footprinting agent o-CuP to probe binary complexes formed between E54 and homoduplex S.meliloti nifH promoter DNA (Table 1), and the ternary E54 complexes formed with the AAA+ activator PspF lacking its C-terminal DNA-binding domain (PspFHTH, amino acids 1–292) (14,25). To form a stable ternary complex between E54 and PspFHTH we used ADP·AlFx, a transition state analogue of ATP that binds to activators of E54 and allows stable binding of PspFHTH to the 54 protein (25). We wanted to learn if E54–DNA interactions within the closed promoter complex were changed by stable binding of the activator. Any changes would provide evidence for activator promoting changes in DNA interactions that could be associated with the nucleotide-dependent stimulation of open complex formation, but which could occur prior to ADP or Pi release. Complexes were formed between the S.meliloti nifH homoduplex promoter probe and E54 in the absence or presence of PspFHTH and ADP·AlFx. DNA footprinting reactions were then carried out and the DNA complexes resolved on native polyacrylamide gels. Subsequently, the DNA cleavage patterns were examined following denaturing gel electrophoresis of the promoter DNA recovered from native gels. Figure 1A shows a native gel indicating the position of binary E54 complexes and the ternary E54–PspFHTH:ADP·AlFx complex from which o-CuP-treated S.meliloti nifH promoter DNA (with end-labelled template strand) was isolated. Figure 1B shows the DNA cutting patterns from the isolated complexes and free DNA seen in Figure 1A. Compared to promoter DNA not exposed to the E54, the E54-bound DNA shows a clear protection from cutting across the promoter region and a strong extra reactivity just downstream of the GC consensus promoter sequence . The extra reactivity (at position –11) corresponds to the position of the native DNA fork junction structure (8,14). In ternary complexes that contain both E54 and PspFHTH:ADP·AlFx the extra reactivity to o-CuP at position –11 is not seen, but it is present in those binary promoter complexes that contain E54 and which were isolated from the same reaction in which the PspFHTH and ADP·AlFx were present (Figure 1B, compare lanes 5b with 5a). These unstable complexes are potentially unstable binary complexes that PspFHTH is unable to bind or modify. Thus, changed cutting patterns are clearly associated with the stable ternary complex.
Figure 1. Changed promoter interactions within ternary E54–PspFHTH:ADP·AlFx promoter complexes. (A) Native gel showing the position of binary E54 complexes (a) and the ternary E54 complex (b) formed by the addition of PspFHTH:ADP·AlFx using the S.meliloti nifH homoduplex promoter probe (see Table 1) following treatment with the footprinting reagent o-CuP, (B) template strand, and (C) non-template strand, o-CuP footprints from unbound DNA or bound promoter complexes (a) and (b) shown in (A). In (B), the position of the native DNA fork junction structure (at –11) and extra o-CuP reactivity at positions –14 and –13 are indicated by an arrow and unfilled arrowheads, respectively. Reactions contain 100 nM promoter probe, 250 nM E54 , 20 μM PspFHTH and 100 ng/μl salmon sperm DNA.
Using homoduplex S.meliloti nifH promoter DNA with the non-template strand end-labelled, complementary results were obtained (Figure 1C). Again the extra fork junction reactivity towards o-CuP seen in DNA bound by the E54 was absent when the PspFHTH:ADP·AlFx was bound to the E54 (Figure 1C, compare lanes 5a and 5b respectively). Controls showed that ADP·AlFx in the absence of PspFHTH did not change the patterns of o-CuP reactivity across the E54-binding region (Figure 1B and C, lane 6a) and that PspFHTH alone did not change the E54 footprint (Figure 1B and C, lane 4a). Further, core RNAP (E) did not yield protection or extra cutting regardless of the presence of ADP·AlFx (Figure 1B and C, compare lanes 7 and 8). Since the footprinting reactions contained salmon sperm DNA as non-specific competitor DNA and PspFHTH:ADP·AlFx alone did not cause changes in cutting patterns under these conditions (Figure 1B and C, compare lanes 2 and 9), we conclude that the PspFHTH:ADP·AlFx causes an altered interaction of o-CuP with promoter DNA in ternary E54 complexes at the place where the fork junction structure is maintained. Attempts to probe the conformation of the A:T base-pair 3' to the 5' GC sequence at –12 using KMnO4 or diethylpyrocarbonate failed because complexes containing PspFHTH:ADP·AlFx were largely destroyed under reaction conditions. Clearly the presence of the activator with ADP·AlFx bound causes changes, in o-CuP reactivity within the closed complex, near the GC promoter element, but does not distinguish between the PspFHTH:ADP·AlFx causing protection of the –12 proximal sequences from cutting or causing a change in the structure of DNA in the –12 region to account for the absence of hyper-reactivity in the ternary complexes.
Further examination of the DNA footprints shown in Figure 1 reveals two other interesting features. First with the ternary complexes that contain the E54 and the PspFHTH:ADP·AlFx (Figure 1B and C, lane 5b), the o-CuP DNA footprint extends further downstream to about –7 compared to binary E54 complexes (Figure 1B and C, lanes 3a, 4a, 5a and 6a), suggesting that interactions between the E54 and PspFHTH:ADP·AlFx increases interaction with the DNA that is to be melted. DNase I footprints conducted under the same assay conditions used in the o-CuP reactions described above have also shown extended footprints towards the transcription start in E54–PspFHTH:ADP·AlFx ternary promoter complexes . Isomerized binary 54–DNA complexes that are activator and NTP hydrolysis-dependent also produce extended DNase I footprints (19), suggesting that the interaction with the promoter DNA which is melted out in isomerized binary complexes (20) has also occurred in ternary E54–PspFHTH:ADP·AlFx complexes, even though no direct evidence for DNA melting in the ternary complex has been demonstrated (described in the later sections). These results provide further evidence that the bound activator leads either directly or indirectly to changed interactions with promoter DNA downstream of the promoter GC element. Second, in the unbound promoter DNA with the template strand end-labelled from the reactions containing E54 (Figure 1B, lanes 3–6) extra reactivity is seen at –14 and –13 (corresponding to the G and C consensus promoter residues respectively) but not when the non-template strand is labelled (Figure 1C, lanes 3–6). The enrichment of specifically cut template strand sequences in the free DNA suggests that these sequences are attacked by o-CuP when E54 binds promoter DNA to form complexes that do not survive native gel electrophoresis. The extra reactivity at –14 and –13 is unchanged by PspFHTH:ADP·AlFx (Figure 1B, compare lanes 3 and 5), in marked contrast to the –11 reactivity (Figure 1B, compare lanes 3a and 5b), implying that one binary complex is specifically changed in an activator plus ADP·AlFx-dependent way, whereas the unstable binary complex is not.
ADP·AlFx bound to activator changes interactions between promoter DNA and 54
Direct binding of PspFHTH:ADP·AlFx to 54 independent of core RNAP has been demonstrated (25). To determine the extent to which the footprints of E54 bound by the PspFHTH:ADP·AlFx complex were contributed to by 54–PspFHTH:ADP·AlFx, we conducted o-CuP footprints of 54 and its complex with PspFHTH:ADP·AlFx on S.meliloti nifH homoduplex DNA. As shown in Figure 2, compared to the binary 54–DNA complex (lane 3), the ternary 54–DNA complex bound with the ADP·AlFx form of the activator (lane 4a) gives additional protection close to the fork junction region, and some extra reactivity located just downstream of the fork junction region. These patterns of cutting are specific since the promoter DNA isolated from the binary 54–DNA complex isolated from the same binding reaction does not show the pattern of cutting by o-CuP seen in the ternary complex (Figure 2, compare lanes 4b with 4a). Rather the cutting pattern is very close to that seen in reactions that just contain 54 and promoter DNA (Figure 2, compare lanes 4b with 3). Overall results confirmed that the interaction of 54 with PspFHTH:ADP·AlFx results in clear changes in DNA footprints across promoter sequences associated with the fork junction structure found in the E54–promoter DNA complex (Figure 2, compare lanes 4a and 5 respectively). It seems that 54–PspFHTH:ADP·AlFx interactions (Figure 2, lane 4a) mediate the changes in cutting patterns seen with the ternary E54–PspFHTH:ADP·AlFx complex (Figure 1B and 1C, lane 5b).
Figure 2. The o-CuP DNA footprints of ternary 54–PspFHTH:ADP·AlFx promoter complexes. Reactions contain 100 nM S.meliloti nifH homoduplex promoter probe (template strand end-labelled), 2 μM 54 or 250 nM E54, 20 μM PspFHTH and 100 ng/μl salmon sperm DNA. Increased reactivity (filled circles) and decreased reactivity (unfilled circles) to o-CuP in the ternary 54–PspFHTH:ADP·AlFx promoter complex is indicated (lane 4a) as well as the increased reactivity corresponding to the DNA fork junction structure seen in binary E54 complexes (arrowed, lane 5). Lane 4b shows the footprint of the binary 54–DNA complex formed in the same reaction (lane 4a) containing PspFHTH:ADP·AlFx.
Ternary holoenzyme complexes have altered –6, +1 interactions
Nucleotide hydrolysis by the activator is needed for 54 and E54 to melt promoter DNA (19,20,32). To help determine whether ternary 54–PspFHTH:ADP·AlFx promoter complexes might contain melted DNA, we used KMnO4 to probe for single-stranded DNA. Although we demonstrated that the binary E54 promoter complexes were closed complexes the ternary E54–PspFHTH:ADP·AlFx promoter complexes were largely destroyed by KMnO4 treatment. Therefore, we used o-CuP footprinting to see if the PspFHTH:ADP·AlFx-dependent 54 ternary complex could be distinguished from a nucleotide-dependent isomerized binary 54 complex in which promoter melting has been demonstrated (20). A heteroduplex promoter DNA probe based on the E.coli glnHp2 m-12 promoter (33), which creates the artificial fork junction next to the GC required for 54 isomerization, was employed since clear DNA melting is evident at this promoter . The native gel shown in Figure 3A indicates the positions of the binary (54–DNA), nucleotide-dependent isomerized 54–DNA (ss–DNA), and E54–DNA and ternary PspFHTH:ADP·AlFx-dependent (54–DNA and E54–DNA) complexes from which o-CuP-treated heteroduplex promoter DNA (template strand end-labelled) was isolated. In marked contrast to the different o-CuP footprints of binary 54–DNA and E54–DNA complexes seen with the homoduplex probe (Figure 2, lanes 3 and 5 respectively), with heteroduplex DNA (E.coli glnHp2) the binary 54 and E54 footprints are indistinguishable (Figure 3B, compare lanes 3a and 6a). This indicates that the heteroduplex region largely fulfils the role of core RNA polymerase in creating the fork junction. Comparing isomerized 54 (Figure 3B, lane 4b) and ternary 54–PspFHTH:ADP·AlFx (Figure 3B, lane 5b) complexes indicates very close similarities in DNA structure within both complexes. The o-CuP footprints (Figure 3B) also clearly show that in all the binary and ternary complexes the o-CuP reactivity around the region of heteroduplex was greatly diminished compared to the free DNA (Figure 3B, lane 2), suggesting a close interaction between 54 and the DNA fork junction structure in each complex. Comparing E54–DNA (Figure 3B, lanes 7a and 8a) to the ternary ADP·AlFx-dependent complex (Figure 3B, lane 8b) shows that in the E54–PspFHTH:ADP·AlFx complex there is some reduced o-CuP cutting around –6, but markedly at +1. Clearly in 54 and E54 complexes with PspFHTH:ADP·AlFx, the interactions with DNA seem different around the transcription start site (+1) (Figure 3B, compare lanes 4b and 8b), suggesting that this is core RNA polymerase-dependent.
Figure 3. o-CuP DNA footprints of ternary ADP·AlFx-dependent and binary nucleotide-dependent isomerized 54 complexes formed on heteroduplex promoter DNA. (A) Native gel showing the position of binary and ternary complexes from which o-CuP-treated promoter DNA was isolated. (B) o-CuP footprints of DNA isolated from complexes shown in (A). E.coli glnHp2 heteroduplex promoter DNA (see Table 1) was at 100 nM, 54 1 μM, E54 was at 100 nM, PspFHTH at 20 μM and salmon sperm DNA at 100 ng/μl. Isomerized 54 binary complexes were formed in the presence of GTP (4 mM) as the hydrolysable nucleotide. Filled arrowheads indicate the position of the region of heteroduplex that is protected in ternary and binary 54 complexes formed on heteroduplex promoter DNA.
DMS footprints reveal conserved promoter contacts in ternary complexes
Conserved G:C base pairs characterize 54-binding sites, with one set adjacent to the –12 proximal DNA fork junction structure formed in E54 closed promoter complexes, the other set at –24. We conducted DNA footprints using DMS to measure the interactions of 54 and E54 in the absence and presence of ADP·AlFx and PspFHTH to learn if PspFHTH:ADP·AlFx changed the specific sets of 54–DNA interactions involving conserved G:C base pairs at the homoduplex S.meliloti nifH promoter (see Table 1). Methylation protection of the guanines was measured. DMS footprints shown in Figure 4 confirm that the consensus G residues (G-14, G-25 and G-26) of the promoter were protected from DMS attack, as expected in binary 54 and E54 complexes (compare lanes 3b, 3f and 5b with 5f, respectively). Comparison of the 54 and E54 binary complex footprints (Figure 4, lanes 3b and 5b) to those of the 54 and E54 ternary complexes (Figure 4, lanes 4b and 6b respectively) formed with PspFHTH:ADP·AlFx did not show any major changes in the patterns of DMS reactivity. A control with just PspFHTH:ADP·AlFx and promoter DNA (Figure 4, lane 7f) showed that the methylation protection observed was dependent on the presence of 54. To learn if the guanine-specific DNA footprints might usefully distinguish between the ternary 54–DNA complex with bound PspFHTH:ADP·AlFx and the nucleotide-dependent isomerized binary 54 complex, DMS footprints were conducted on S.meliloti nifH DNA with a short region of heteroduplex next to the promoter GC at –12 necessary for efficient isomerization (see Table 1). Comparisons of the DNA footprints from binary 54–DNA and isomerized 54–DNA complexes and the ternary 54–DNA complex showed that conserved guanines were protected from methylation by DMS and that the footprints were indistinguishable (data not shown). Binary 54–RNAP and ternary E54–PspFHTH:ADP·AlFx complexes formed with the heteroduplex promoter DNA also showed similar DMS footprints with conserved guanines being protected from methylation by DMS (data not shown). The DMS footprinting results indicate that key 54–DNA promoter contacts are unchanged by binding activator in the transition state for ATP hydrolysis, suggesting that changes in o-CuP reactivity at –11 (Figure 1B and C) arise when contacts at the major promoter specificity sequences are being maintained.
Figure 4. DMS footprints on homoduplex S.meliloti nifH promoter DNA. Promoter DNA was isolated from unbound (lanes marked with ‘f’) and bound (lanes marked with ‘b’) binary and ternary 54 and E54 complexes following the separation on a native polyacrylamide gel (data not shown). Reactions included 100 nM homoduplex S.meliloti nifH promoter DNA (non-template strand end-labelled), 2 μM 54, 250 nM E54, 20 μM PspFHTH and 100 ng/μl salmon sperm DNA. The position of consensus guanines (G-14, G-25 and G-26) that are protected from methylation in binary and ternary promoter complexes is indicated.
Interactions between activator and DNA
To examine whether there might be a direct activator–DNA interaction within the ternary E54–PspFHTH:ADP·AlFx promoter complexes described above, we conducted o-CuP DNA footprinting reactions using the heteroduplex E.coli glnHp2 DNA template (see Table 1) with increasing amounts of PspFHTH in the absence or presence of ADP·AlFx. The results shown in Figure 5A indicate that increasing the concentration of PspFHTH does lead to a reduced o-CuP reactivity across, and downstream of, the region of heteroduplex implying that there is potential for a DNA contact made by the AAA+ domain of PspFHTH within the ternary complexes. The presence of ADP·AlFx did not appear to have any effect on PspFHTH binding (Figure 5A, compare lanes 4, 6, 8 and 10 with lanes 3, 5, 7 and 9, respectively), and the protein storage buffer did not cause loss of cutting by o-CuP (data not shown). Reduced o-CuP reactivity across, and downstream of the region of heteroduplex on the non-template strand was also observed with PspFHTH and the addition of ADP·AlFx had no effect on PspFHTH binding (data not shown). Forms of PspFHTH that carried different N-terminal tags, or had the His-tag removed also gave a footprint characterized by reduced DNA reactivity downstream of the –12 GC consensus promoter sequence (data not shown) providing evidence that the tags on PspF sequences do not contribute to the pattern of o-CuP cutting seen in the footprinting assays. The reduced DNA reactivity was also seen on the template and non-template strands with a construct comprising amino acids 1–275 of PspF (PspF1–275), the minimum region of PspF corresponding to the AAA+ domain (Figure 5B, left-hand panel and data not shown). Therefore, the o-CuP footprint is not due to sequences that link the AAA+ domain of PspF to its DNA-binding domain. Similar results were obtained when we replaced the heteroduplex E.coli glnHp2 DNA with the S.meliloti nifH heteroduplex (see Table 1) promoter probe, suggesting that PspF1–275 may contact 54-dependent promoters in general (data not shown). We repeated the o-CuP footprinting reactions shown in Figure 5A but replaced the heteroduplex E.coli glnHp2 promoter probe with homoduplex E.coli glnHp2 promoter DNA (see Table 1). The homoduplex E.coli glnHp2 probe alone displayed strong o-CuP reactivity between about –6 and –9, indicating that a particular minor groove geometry there expected from its A:T richness (Figure 5C, compare lanes 1 and 2). An interaction of PspFHTH with the E.coli glnHp2 homoduplex DNA was also evident as judged by reduced o-CuP reactivity downstream of the –12 GC (see Figure 5C), indicating that the region of heteroduplex was not required for the binding of PspFHTH.
Figure 5. Activator interacts with heteroduplex and homoduplex E.coli glnHp2 promoter DNA. (A) o-CuP DNA footprints resulting from the addition of increasing amounts of PspFHTH (2, 5, 10 and 20 μM) to 100 nM heteroduplex E.coli glnHp2 promoter DNA (see Table 1) in the absence or presence of ADP·AlFx. (B) o-CuP footprints with pspF1–275 at 2, 5, 10 and 20 μM. (C) o-CuP footprints as in (A), except that homoduplex E.coli GlnHp2 promoter probe (see Table 1) was used. Salmon sperm DNA was not included in the binding reactions. The region of heteroduplex is indicated by filled arrowheads, and the promoter region showing reduced o-CuP reactivity in the presence of activator is indicated by a solid line. Lane ‘m’, size marker generated by end-labelling oligonucleotides.
To examine whether the DNA footprints due to the presence of PspF alone were specific to certain PspF residues, increasing amounts (5–20 μM) of mutant PspF1–275 (the minimal AAA + domain of PspF) proteins (25,28,31) were added to the heteroduplex E.coli glnHp2 promoter probe and the DNA examined by o-CuP footprinting. The results (data not shown) show that the single amino acid substitution mutant PspF1–275 proteins tested (K42A, T86A, T86S, D107A, R122A, T148A, N149A, R162A, R168A, R227A, K230A and N231A) with defects in the GAFTGA, Walker A, Walker B motifs and so-called Sensor regions (22) are not significantly defective (at most a 2-fold reduction in binding for the D107A mutant) in promoter interaction. Similar to PspF1–275, the mutant PspF1–275 proteins showed similar protection downstream of the –12 GC consensus promoter sequence from o-CuP cutting, suggesting that the substituted amino acid residues do not contribute to the DNA-binding activity of the activator.
In order to determine whether the o-CuP footprint, we observed with the AAA+ domain of PspF was evident with another protein we purified the AAA+ domain from the 54-dependent activator, C4-dicarboxylic acid transport protein D (DctD) from S.meliloti (see Materials and Methods). The results using the heteroduplex E.coli glnHp2 promoter probe with the template strand end-labeled (Figure 5B, right-hand panel) or non-template strand end-labeled (data not shown) clearly show that the DctD AAA+ domain gives a similar o-CuP footprint to that of PspF1–275 (Figure 5B, left-hand panel and data not shown) and protects the DNA from o-CuP cutting, across and downstream from the region of heteroduplex. The presence of ADP·AlFx had no effect upon DctD binding (data not shown).
When the binding reactions between activator and DNA in the absence or presence of ADP·AlFx (see Materials and Methods) were run on native polyacrylamide gels no strong clear band shifted binary activator–DNA complexes were seen (see Figure 1A, lane 10 and Figure 3A, lane 9 and data not shown), suggesting that the DNA binding that gave rise to the o-CuP footprints was very unstable.
A composite 54, PspFHTH DNA footprint
Consistent with the binding of PspFHTH to promoter DNA being weak, titration of salmon sperm DNA into reactions containing promoter probe and PspFHTH resulted in a gradual loss of protection, with no protection being seen above 100 ng/μl salmon sperm DNA (data not shown). Recall that salmon sperm DNA was included in the earlier o-CuP and DMS footprints (Figures 1–4), and can explain why the control reactions in those experiments (i.e. probe plus PspFHTH) do not reveal a direct interaction of PspFHTH with DNA. As shown in Figure 6, inclusion of 54 in binding reactions with PspFHTH and the E.coli glnHp2 promoter probe (template strand end-labelled) in the absence of salmon sperm DNA (lane 6) resulted in an o-CuP footprint that was apparently a composite of the 54 footprint (lane 3) and that due to PspFHTH alone (lane 8). Under conditions where the 54 protein stably bound to the activator through the inclusion of ADP·AlFx a composite o-CuP footprint was again evident (Figure 6, lane 7). A similar composite footprint was also observed in the nucleotide-dependent isomerized binary 54–DNA complex (Figure 6, lane 4). Although composite footprints are observed independent of the nucleotide (lanes 4, 6 and 7), only PspFHTH in the presence of ADP·AlFx forms stable ternary complexes, suggesting different structural states that are nucleotide-dependent. Addition of salmon sperm DNA diminished the protection associated with PspFHTH alone, but the protection due to 54 persisted (see Figure 3B), suggesting that the interaction between the fork junction DNA and the activator is weak, and if occurring in ternary promoter complexes, would have to be stabilized by interactions dependent upon 54 and E54.
Figure 6. ‘Composite’ o-CuP DNA footprints of ternary ADP·AlFx-dependent and binary nucleotide-dependent isomerized 54 complexes formed on heteroduplex E.coli glnHp2 promoter DNA in the absence of salmon sperm DNA. Reactions were essentially as in Figure 3 but salmon sperm DNA was not included in the binding reactions. Filled arrowheads indicate the position of the region of heteroduplex that is protected in ternary and binary 54 complexes formed on heteroduplex promoter DNA.
DNA cross-linking to PspF1–275 within a ternary complex
To further investigate possible activator interactions with promoter DNA, we used photoreactive DNA probes (29). The promoter probes were based on the heteroduplex S.meliloti nifH promoter (see Table 1) with a photoactivatable cross link (azidophenacyl bromide) tethered to the DNA via a uniquely positioned phosphorothioate residue either upstream or downstream of the region of heteroduplex (i.e. between bases, –15/–14, –14/–13, –13/–12, –12/–11, –11/–10 and –7/–6) and so including positions that are protected in o-CuP DNA footprints described above, with the phosphorothiolated strand 32P-end-labelled.
Initially, we incubated 54 alone and 54 in the presence of PspF1–275:ADP·AlFx using a DNA probe with the phosphorothioate residue located between bases –7 and –6 on the template strand. The binding reactions were then irradiated prior to loading onto a native polyacrylamide gel (Figure 7, left-hand panel). Binary 54–DNA and ternary 54–PspF1–275:ADP·AlFx complexes clearly formed with the cross-linking probe (Figure 7, labelled ‘b’ in lane 2, and ‘c’ in lane 3, respectively). Control reactions with either PspF1–275 alone (Figure 7, lane 4) or PspF1–275 in the presence of ADP·AlFx (data not shown) revealed no gel-shifted complexes following UV cross-linking. The binary and ternary complexes and unbound DNA resolved on the native gel (Figure 7, lanes 1–4, labelled a–d) were eluted into 1x Laemmli sample buffer (see Materials and Methods) and then further analysed by SDS–PAGE (Figure 7, right-hand panel, lanes a–d). In marked contrast to the binary 54–DNA complex, analysis of the isolated ternary 54–PspF1–275:ADP·AlFx complex on SDS–PAGE showed the presence of an additional protein–DNA cross-linked complex (Figure 7, compare lanes b and c). This band is PspF1–275 cross-linked to the DNA probe (PspF1–275–DNA) as established by using different molecular weight forms of PspF (Burrows et al., manuscript in preparation). The PspF1–275–DNA cross-linked complex (Figure 7, lane c) was only observed when both 54 and ADP·AlFx were present with PspF1–275. PspF1–275 cross-linked to DNA in the presence of RNAP holoenzyme, demonstrating PspF1–275 proximity to DNA in a complex competent for transcriptional activation (30). Also, similar results were obtained with all the cross-linking DNA probes tested (outlined above) with the phosphorothioate residue located on either the template or non-template strand (data not shown). The cross-linking results suggest that within the ternary 54–PspF1–275:ADP·AlFx complex PspF1–275 is proximal to regions of the DNA either side of the region of heteroduplex.
Figure 7. DNA-crosslinking to PspF1–275 is dependent on 54 and ADP·AlFx. Reactions contain a modified heteroduplex S.meliloti nifH promoter probe (100 nM) based on the heteroduplex probe shown in Table 1, but with a photoactivatable cross link positioned between bases –7 and –6 on the template strand. Where indicated PspF1–275 (20 μM), 54 (1 μM) and ADP·AlFx (0.2 mM) are included in the DNA-binding assay. Reactions were then irradiated for 5 min and loaded onto a 4.5% native polyacrylamide gel (left-hand panel). The binary 54–DNA (b, lane 2) and ternary 54–PspF1–275:ADP·AlFx (c, lane 3) complexes, the free DNA (a, lane 1) and the free DNA from the reaction containing PspF1–275 (d, lane 4) were further analysed by SDS–PAGE (right-hand panel).
DNA stimulates the activity of the PspF AAA+ domain, suggesting a functional role for DNA
To help determine a functional relevance for the activator interaction with DNA, we measured the ATPase activity of PspF1–275 in the absence or presence of DNA. In the presence of either homoduplex or heteroduplex E.coli glnHp2 promoter DNA (see Table 1), PspF1–275 repeatedly showed a modest but clear increase in ATPase activity compared to PspF1–275 alone (Figure 8). The increase in activity was not due to ATPase activity of the DNA alone (Figure 8 and data not shown). Next, to determine if the stimulation is promoter sequence specific, we carried out the same experiments in the presence of salmon sperm DNA. The same level of ATPase stimulation was observed with salmon sperm DNA, demonstrating that the stimulation of PspF1–275 is DNA-dependent but not obviously DNA sequence-dependent. This is in agreement with our DNA foot-printing experiments where salmon sperm DNA quenches DNA protection of the promoter DNA. These observations suggest that relatively weak and unspecific interactions between the AAA+ domain of PspF (PspF1–275) and the promoter DNA in the ternary E54–DNA complex may play a functional role by contributing to productive interactions within a DNA-dependent architecture of the initiation complex. Although PspF1–275–DNA interactions are unspecific, 54–PspF1–275 interactions could direct PspF proximal to the promoter sequence. We considered that DNA could increase the ATPase activity of PspF1–275 by promoting higher order oligomer formation of PspF1–275 or by increasing the activity of the catalytic site of PspF1–275 subunits. Evidence for an effect upon higher order oligomer formation comes from the finding that stimulation by DNA is dependent upon PspF concentration in the range where PspF1–275 oligomerizes (31) and is independent of ATP concentration (data not shown).
DISCUSSION
The activators of E54 are mechanochemical ATPases that use ATP hydrolysis to promote DNA strand opening and stimulate the associated changes in 54 structure needed for open complex formation (6,11,19,22,25). The ways in which nucleotide binding and its hydrolysis used for open complex formation are not well understood, but the effects of ADP·AlFx presented here begin to address this issue. Several lines of evidence, notably the use of DNA fork junction promoter probes and related heteroduplex probes, have shown that the E54 makes use of the double-strand–single-strand junction to limit DNA opening (8,11,12,19,20,23,26). Changed E54 interactions at the fork junction appear to be required to allow open complexes to form (19,34) and our new data strongly support and extend this view. Variations in promoter sequences suggest that a range of natural DNA fork junctions will exist. The functional state of the activator required for stable binding to the E54 was created by interaction with ADP·AlFx, an analogue of ATP at the transition state for hydrolysis. The N-terminal Region I of 54 is required for creating the fork junction structure and activator binds directly to Region I, suggesting that activator changes interactions between the fork junction structure and 54 in a nucleotide-dependent manner (25). A comparison (Figure 1B) of bound and unbound DNA exposed to E54 showed that the effects of PspFHTH:ADP·AlFx were restricted to complexes resolved in native gels, and that the effects of PspFHTH:ADP·AlFx upon unstable E54 promoter DNA complexes were not evident. Consideration of previous o-CuP footprinting (14) carried out in solution and without resolving complexes indicate that these footprints in fact arise from a mixture of stable complexes (o-CuP signal in the –11 region) and unstable complexes (o-CuP signal in the –14 region).
The changed DNA interactions at the fork junction structure (which may be entirely indirect and through 54, or may involve a DNA contact by the AAA+ domain of PspF) suggest that a part of the DNA opening pathway involves an activator interaction with promoter DNA, potentially either side of the fork junction structure. The PspF AAA+ domain showed clear protection around the region of heteroduplex present in the DNA probes used in this work. A number of conserved residues in PspF appeared not to contribute to the DNA binding shown by the minimal AAA+ domain, implying that interactions of PspF1–275 with promoter DNA will not greatly change the activities of PspF associated with these conserved amino acids. Some other AAA+ proteins notably those related to RuvB that remodel DNA and nucleoprotein complexes are thought to make direct interactions with the DNA component of their targets. In these cases, a DNA-dependent stimulation of the ATPase function has been reported, and we have also observed this (see Figure 8) for the minimal AAA+ domain of PspF (PspF1–275) and for PspFHTH (data not shown) in the presence of promoter DNA and non-specific DNA. This supports the fact that weak but direct PspF1–275 contacts to the DNA occur and are significant for the nucleotide-dependent formation of open promoter complexes. Direct evidence for AAA+ domain–DNA interactions comes from the recently published structures from replication factor C of the clamp loader complex, involved in DNA replication (38,39). In our current model, specific DNA binding is mediated by the PspF HTH motif to its upstream enhancer site. However, in the ternary E54–PspF1–275:ADP·AlFx–DNA complex, PspF1–275 (the AAA+ domain) is proximal to the promoter DNA where it could contribute to the architecture of the ternary complex by interacting with DNA sequences across and downstream of the –12 GC. In the context of the full-length PspF protein, the downstream promoter interactions by the AAA+ domain may be enhanced by the specific DNA-binding activity of the HTH motif. upstream. Qualitative aspects of the footprint displayed by PspF1–275 (Figure 5B) were independent of any functional state we examined, including the ADP·AlFx bound state, the ADP bound state and 54 Region I bound state (data not shown).
A related architecture between RuvB (35) and PspF is predicted (30,36) and some relatedness in where these proteins interact with their targets is expected. Interactions between the Region I of 54, the fork junction DNA and PspF might therefore be direct, and help to modify the ATPase domain structure and activity to enable remodelling of the closed promoter complex. RuvA could interact with RuvB as 54 interacts with its activators. RuvB and PspF could contact their respective DNA targets.
A common core RNA polymerase enzyme is used by the enhancer-independent class of sigma factors. It would appear that the special features of 54-dependent transcription relate closely to the activator targeting an unusually conformationally stable fork junction containing complex by making contact with the DNA and protein components of the closed complex. Some of the 54–DNA interactions that activator changes seem to have a modest energetic cost, as evidenced by the action of the ATPS bound activator (26). Others, notably the DNA opening per se seem to correlate to a full ATP hydrolysis event and appear to have a higher energetic cost. The sensing of the gamma phosphate of the ATP is implicated as critically changing the functionality of the activator (22), a common theme for AAA+ proteins where nucleotide binding and hydrolysis control the binding interactions needed for substrate remodelling. For activation of promoters by the catabolite activating protein CAP, remodelling of promoter complexes occurs, but uses an activator–RNA polymerase contact upstream of the main promoter sequences (37). The use of DNA looping and enhancers appears to enable the 54 activators to make a contact just downstream of the main promoter elements, near or at the place where the fork junction exists in closed complexes. It now appears that core promoter DNA is contacted by the AAA+ domain of the activator, in addition to the promoter contacts made by 54.
ACKNOWLEDGEMENTS
We thank P.Bordes for PspF proteins, P.Ray for 54 protein and T. Hoover for plasmid pHX187 encoding the S.meliloti DctD AAA+ domain. The DNA probes used in cross-linking experiments were designed and prepared by S. Wigneshweraraj and P.Burrows, whose help is gratefully acknowledged. We also thank S.Wigneshweraraj and G.Jovanovic for comments on the manuscript. Work was supported by a Wellcome Trust project grant to M.B.
REFERENCES
Ebright,R.H. ( (2000) ) RNA polymerase: structural similarities between bacterial RNA polymerase and eukaryotic RNA polymerase II. J. Mol. Biol., , 304, , 687–698.
Cramer,P. ( (2002) ) Multisubunit RNA polymerases. Curr. Opin. Struct. Biol., , 12, , 89–97.
Gross,C.A., Chan,C., Dombroski,A., Gruber,T., Sharp,M., Tupy,J. and Young,B. ( (1998) ) The functional and regulatory roles of sigma factors in transcription. Cold Spring Harb. Symp. Quant. Biol., , 63, , 141–155.
Burgess,R.R. and Anthony,L. ( (2001) ) How sigma docks to RNA polymerase and what sigma does. Curr. Opin. Microbiol., , 2, , 126–131.
Studholme,D.J. and Buck,M. ( (2000) ) The alternative sigma factor sigma(28) of the extreme thermophile Aquifex aeolicus restores motility to an Escherichia coli fliA mutant. FEMS Microbiol. Lett., , 191, , 103–107.
Buck,M., Gallegos,M.-T., Studholme,D.J., Guo,Y. and Gralla,J.D. ( (2000) ) The bacterial enhancer-dependent 54 (N) transcription factor. J. Bacteriol., , 15, , 4129–4136.
Xu,H. and Hoover,T.R. ( (2001) ) Transcriptional regulation at a distance in bacteria. Curr. Opin. Microbiol., , 4, , 138–144.
Guo,Y., Wang,L. and Gralla,J.D. ( (1999) ) A fork junction DNA–protein switch that controls promoter melting by the bacterial enhancer-dependent sigma factor. EMBO J., , 18, , 3736–3745.
Guo,Y. and Gralla,J.D. ( (1998) ) Promoter opening via a DNA fork junction binding activity. Proc. Natl Acad. Sci. USA, , 95, , 11655–11660.
Chaney,M. and Buck,M. ( (1999) ) The 54 DNA-binding domain includes a determinant of enhancer responsiveness. Mol. Microbiol., , 33, , 1200–1209.
Guo,Y., Lew,C.M. and Gralla,J.D. ( (2000) ) Promoter opening by sigma(54) and sigma(70) RNA polymerases: sigma factor-directed alterations in the mechanism and tightness of control. Genes Dev., , 14, , 2242–2255.
Wedel,A. and Kustu,S. ( (1995) ) The bacterial enhancer-binding protein NTRC is a molecular machine: ATP hydrolysis is coupled to transcriptional activation. Genes Dev., , 9, , 2042–2052.
Cannon,W., Gallegos,M.-T., Casaz,P. and Buck,M. ( (1999) ) Amino-terminal sequences of N (54) inhibit RNA polymerase isomerization. Genes Dev., , 13, , 357–370.
Morris,L., Cannon,W., Claverie-Martin,F., Austin,S. and Buck,M. ( (1994) ) DNA distortion and nucleation of local DNA unwinding within 54 (N) holoenzyme closed promoter complexes. J. Biol. Chem., , 269, , 11563–11571.
Wigneshweraraj,S.R., Konstantin,K., Konstantin,S. and Buck,M. ( (2003) ) Multiple roles of the RNA polymerase ? subunit flap domain in54-dependent transcription. J. Biol. Chem., , 278, , 3455–3465.
Murakami,K.S., Masuda,S. and Darst,S.A. ( (2002) ) Structural basis of transcription initiation: RNA polymerase holoenzyme at 4 ? resolution. Science, , 296, , 1280–1284.
Murakami,K.S., Masuda,S., Campbell,E.A., Muzzin,O. and Darst,S.A. ( (2002) ) Structural basis of transcription initiation: an RNA polymerase holoenzyme–DNA complex. Science, , 296, , 1285–1290.
Tsujikawa,L., Tsodikov,O.V. and deHaseth,P.L. ( (2002) ) Interaction of RNA polymerase with forked DNA: evidence for two kinetically significant intermediates on the pathway to the final complex. Proc. Natl Acad. Sci. USA, , 99, , 3493–3498.
Cannon,W.V., Gallegos,M.-T. and Buck,M. ( (2000) ) Isomerization of a binary sigma–promoter DNA complex by transcription activators. Nature Struct. Biol., , 7, , 594–601.
Cannon,W., Gallaegos,M.-T. and Buck,M. ( (2001) ) DNA melting within a binary sigma(54)–promoter DNA complex. J. Biol. Chem., , 276, , 386–394.
Rombel,I., North,A., Hwang,I., Wyman,C. and Kustu,S. ( (1998) ) The bacterial enhancer-binding protein NtrC as a molecular machine. Cold Spring Harb. Symp. Quant. Biol., , 63, , 157–66.
Zhang,X., Chaney,M., Wigneshweraraj,S.R., Schumacher,J., Bordes,P., Cannon,W. and Buck,M. ( (2002) ) Mechanochemical ATPases and transcriptional activation. Mol. Microbiol., , 45, , 895–903.
Cannon,W., Wigneshweraraj,S.R. and Buck,M. ( (2002) ) Interactions of regulated and deregulated forms of the sigma54 holoenzyme with heteroduplex promoter DNA. Nucleic Acids Res., , 30, , 886–893.
Wang,L., Guo,Y. and Gralla,J.D. ( (1999) ) Regulation of sigma54-dependent transcription by core promoter sequences: role of –12 region nucleotides. J. Bacteriol., , 181, , 7758–7765.
Chaney,M., Grande,R., Wigneshweraraj,S.R., Cannon,W., Casaz,P., Gallegos,M.T., Schumacher,J., Jones,S., Elderkin,S., Dago,A.E., Morett,E. and Buck,M. ( (2001) ) Binding of transcriptional activators to sigma54 in the presence of the transition state analog ADP–aluminium chloride: insights into activator mechanochemical action. Genes Dev., , 15, , 2282–2294.
Cannon,W., Bordes,P., Wigneshweraraj,S.R. and Buck,M. ( (2003) ) Nucleotide-dependent triggering of RNA polymerase–DNA interactions by an AAA regulator of transcription. J. Biol. Chem., , 278, , 19815–19825.
Jovanovic,G., Rakonjac,J. and Model,P. ( (1999) ) In vivo and in vitro activities of the Escherichia coli sigma54 transcription activator, PspF, and its DNA-binding mutant, PspFDeltaHTH. J. Mol. Biol., , 285, , 469–483.
Bordes,P., Wigneshweraraj,S.R., Schumacher,J., Zhang,X., Chaney,M. and Buck,M. ( (2003) ) The ATP hydrolyzing transcription activator phage shock protein F of Escherichia coli: identifying a surface that binds 54. Proc. Natl Acad. Sci. USA, , 100, , 2278–2283.
Mayer,A.N. and Barany,F. ( (1995) ) Photoaffinity cross-linking of TaqI restriction endonuclease using an aryl azide linked to the phosphate backbone. Gene, , 153, , 1–8.
Burrows,P.C., Serinov,K., Buck,M. and Wigneshweraraj,S.R. ( (2004) ) Re-organisation of an RNA polymerase-promoter DNA complex for DNA melting. EMBO J., , in press.
Schumacher,J., Zhang,X., Jones,S., Bordes,P. and Buck,M. ( (2004) ) ATP-dependent transcriptional activation by bacterial PspF AAA+ protein. J. Mol. Biol., , 338, , 863–875.
Popham,D., Szesto,D., Keener,J. and Kustu,S. ( (1989) ) Function of a bacterial activator protein that binds to transcriptional enhancers. Science, , 243, , 629–635.
Claverie-Martin,F. and Magasanik,B. ( (1992) ) Positive and negative effects of DNA bending on activation of transcription from a distant site. J. Mol. Biol., , 227, , 996–1008.
Wigneshweraraj,S.R., Chaney,M.K., Ishihama,A. and Buck,M. ( (2001) ) Regulatory sequences in sigma 54 localise near the start of DNA melting. J. Mol. Biol., , 306, , 681–701.
Han,Y.W., Iwasaki,H., Miyata,T., Mayanagi,K., Yamada,K., Morikawa,K. and Shinagawa,H. ( (2001) ) A unique beta-hairpin protruding from AAA+ ATPase domain of RuvB motor protein is involved in the interaction with RuvA DNA recognition protein for branch migration of Holliday junctions. J. Biol. Chem., , 276, , 35024–35028.
Lee,S.-Y., De La Torre,A., Yan,D., Kustu,S., Nixon,B.T. and Wemmer,D.E. ( (2003) ) Regulation of the transcriptional activator NtrC1: structural studies of the regulatory and AAA+ ATPase domains. Genes Dev., , 17, , 2552–2563.
Chen,H., Tang,H. and Ebright,R.H. ( (2003) ) Functional interaction between RNA polymerase alpha subunit C-terminal domain and sigma70 in UP-element-and activator-dependent transcription. Mol. Cell, , 11, , 1621–1633.
Miyata,T., Oyama,T., Mayanagi,K., Ishino,S., Ishino,Y. and Morikawa,K. ( (2004) ) The clamp-loading complex for processive DNA replication. Nature Struct. Biol., , 11, , 632–636.
Bowman,G.D., O'Donnell,M. and Kuriyan,J. ( (2004) ) Structural analysis of a eukaryotic sliding DNA clamp–clamp loader complex. Nature, , 429, , 724–730.(W. V. Cannon, J. Schumacher and M. Buck*)