Mutational analysis of σ⁷⁰ region 4 needed for appropriation by the bacteriophage T4 transcription factors AsiA and MotA

Abstract

Transcriptional activation of T4 middle promoters requires σ⁷⁰-containing E. coli RNA polymerase, the T4 activator MotA, and the T4 co-activator AsiA. T4 middle promoters contain the σ⁷⁰ –10 DNA element. However, these promoters lack the σ⁷⁰ –35 element, having instead a MotA box centered at –30, which is bound by MotA. Previous work has indicated that AsiA and MotA interact with region 4 of σ⁷⁰, the C-terminal portion that normally contacts –35 DNA and the β-flap structure in core. AsiA binding prevents the σ⁷⁰/β-flap and σ⁷⁰/-35 DNA interactions, inhibiting transcription from promoters that require a –35 element. To test the importance of residues within σ⁷⁰ region 4 for MotA and AsiA function, we investigated how σ⁷⁰ region 4 mutants interact with AsiA, MotA, and the β-flap and function in transcription assays in vitro. We find that alanine substitutions at residues 584-588 (region 4.2) do not impair the interaction of region 4 with the β-flap or MotA, but they eliminate the interaction with AsiA and prevent AsiA inhibition and MotA/AsiA activation. In contrast, alanine substitutions at 551-552, 554-555 (region 4.1) eliminate the region 4/β-flap interaction, significantly impair the AsiA/σ⁷⁰ interaction, and eliminate AsiA inhibition. However, the 4.1 mutant σ⁷⁰ is still fully competent for activation if both MotA and AsiA are present. A previous NMR structure shows AsiA binding to σ⁷⁰ region 4, dramatically distorting regions 4.1 and 4.2 and indirectly changing the conformation of the MotA interaction site at the σ⁷⁰ C-terminus. Our analyses provide biochemical relevance for the σ⁷⁰ residues identified in the structure, indicate that the interaction of AsiA with σ⁷⁰ region 4.2 is crucial for activation, and support the idea that AsiA binding facilitates an interaction between MotA and the far C-terminus of σ⁷⁰.

Introduction

Prokaryotic RNA polymerase holoenzyme is composed of a core (subunits β, β′, α₂, ω) that has RNA synthesizing activity and a σ factor that imparts DNA specificity ^{1; 2}. The primary σ factor in E. coli, σ⁷⁰, recognizes promoters that are needed for transcription of general housekeeping genes and thus, is responsible for most of the transcription during exponential growth ³. Sequence, structure, and biochemical studies indicate that σ⁷⁰ shares four regions of similarity with primary σ proteins of other prokaryotes ^{1; 4}. When present in RNA polymerase holoenzyme (Eσ⁷⁰), σ⁷⁰ sets the start site for transcription by recognizing various DNA elements, and residues within each of the 4 regions have been shown to interact with sequences in promoter DNA ^{5; 6; 7; 8; 9}. The majority of E. coli promoters have –10 and –35 DNA elements, which are contacted by σ⁷⁰ regions 2 and 4, respectively ^{8; 9}. Specific base recognition of the -35 sequences arises through a DNA-binding helix-turn-helix (H-T-H) motif in region 4.2 (H3-T-H4 in Fig. 1A). In addition, interactions between region 4 and core, particularly a structure called the β-flap, are required for proper positioning of regions 2 and 4 so they can contact the –10 and –35 elements simultaneously ^{10; 11; 12; 13} (Fig. 1A).

Fig. 1.

Region 4 of wild type σ⁷⁰ and mutant σ proteins and the T4 middle promoter P_uvsX. A. The C-terminal amino acid sequence of σ⁷⁰ is shown from residues 540 to 613. Residues in pink contact the -35 element of E. coli promoter DNA ⁸ while residues in blue contact the β-flap tip ¹¹; ¹²; ¹³; ⁴⁵. Every tenth residue is marked with a small vertical line. Above the sequence is shown the locations of regions 4.1 and 4.2 ¹ and the positions of the α-helices H1, H2, and H5 and the helix-turn-helix H3-T-H4 structures observed in region 4 of the primary σ of T. aquaticus complexed with -35 element DNA ⁸. Residues that contact AsiA in the AsiA/σ⁷⁰ region 4 structure ³⁰ are designated with an asterisk and the region of the MotA contact site ²⁰ is indicated. Below are shown the region 4 sequences of the mutant σ proteins, with residues that differ from σ⁷⁰ in red. B. The sequence of P_uvsX from −61 to +9 is shown with the start of transcription indicated as +1. The locations of the σ⁷⁰ –10 element and the MotA box are indicated. Sequences within P_uvsX that match the canonical polymerase elements σ⁷⁰ -10 element (TATAAT); σ⁷⁰ extended -10 element (-15T, -14G); σ⁷⁰ -35 element (TTGACA); and α UP element (AAAa/ta/tTa/tTTTTnnAAAA) are in red.

The presence of a good match to both the -10 and -35 canonical sequences is usually sufficient for Eσ⁷⁰ to recognize a promoter without the aid of additional factors. However, activation factors can be needed when either sequence element deviates significantly from the consensus. Many activators in E. coli work through class I or class II activation mechanisms (reviewed in ^{14; 15}). Class I activators interact with residues in the C-terminal domain (CTD) of the α subunit of polymerase whereas a class II activator may also interact with residues in σ⁷⁰ region 4. MotA/AsiA activation of bacteriophage T4 middle promoters arises through a different mechanism, called σ appropriation (reviewed in ¹⁶). MotA protein is a transcriptional activator ^{17; 18; 19}, whose CTD binds to a MotA box element, centered at -30 in T4 middle promoter DNA, and whose NTD interacts with the far C-terminal portion of σ⁷⁰, ²⁰ (Fig. 1A). AsiA is needed to co-activate transcription from T4 middle promoter ²¹. However, AsiA by itself is also an inhibitor of σ⁷⁰-dependent transcription from -10/-35 promoters ^{22; 23; 24}. Several studies have indicated that AsiA binds tightly to σ⁷⁰ regions 4.1 and 4.2 ^{22; 23; 25; 26; 27; 28; 29; 30; 31; 32} (Fig. 1A). A recent structure of the AsiA/σ⁷⁰ region 4 complex implicated 17 σ⁷⁰ residues as AsiA contacts and indicated that AsiA structurally remodels region 4, converting the DNA–binding H3-T-H4 to a continuous helix ³⁰. It has been suggested that this remodeling could also serve to facilitate the interactions of MotA with the MotA box and with the C-terminus of σ⁷⁰, ^{20; 30}.

Here we have performed mutational analyses to investigate the molecular interface used by σ⁷⁰ with AsiA and MotA, and to ask whether there is biochemical support for σ⁷⁰ residues that were identified in the σ⁷⁰ region 4/AsiA structure. We demonstrate that substitutions of σ⁷⁰ residues in regions 4.1 and 4.2 that interact with AsiA in the AsiA/σ⁷⁰ region 4 structure ³⁰ do indeed impair the interaction of AsiA with σ⁷⁰. We also find that substitutions within region 4.1 that significantly compromise the interaction of AsiA with σ⁷⁰ impair the ability of AsiA to inhibit transcription. However, in the presence of both MotA and AsiA, polymerase containing the 4.1 substitutions remains fully competent for activation. In contrast, substitutions in σ⁷⁰ region 4.2 that eliminate the AsiA/σ⁷⁰ interaction render polymerase immune both to AsiA inhibition and MotA/AsiA activation. Our work indicates that the interaction of AsiA with its contacts in σ⁷⁰ region 4.2 is crucial for σ appropriation by MotA and AsiA and supports a model in which AsiA binding and remodeling of σ⁷⁰ region 4 serve to facilitate the interaction of MotA with σ⁷⁰ and with MotA box DNA.

Results

Rationale

Based on crystal structures using the primary σ factors of T. aquaticus and T. thermophilus and biochemical analyses with σ⁷⁰ mutants (reviewed in ^{2; 33}), specific residues with σ⁷⁰ region 4 are thought to be crucial for transcription from σ⁷⁰-dependent -10/-35 promoters, either because they contact the -35 DNA element or because they interact with the β-flap structure of core. The NMR structure of a complex of AsiA with σ⁷⁰ region 4³⁰ shows that many of these same residues contact AsiA and thus, predicts that these residues may also be important for AsiA inhibition and/or MotA/AsiA activation. In addition, biochemical evidence ²⁰ has suggested that some of the residues within the far C-terminal portion of σ⁷⁰ that normally interact with the β-flap may coincide with residues that contact MotA during MotA/AsiA activation. (These various residues in regions 4.1 and 4.2 and the C-terminal portion of σ⁷⁰ are indicated at the top of Fig. 1A.)

To investigate the functional significance of several of the σ⁷⁰ residues that the structural work predicts could be important for AsiA and MotA function, we used σ⁷⁰ proteins with mutations within regions 4.1 and 4.2 and the C-terminal portion of σ⁷⁰ (listed in the bottom portion of Fig. 1A). Mutations in region 4.1 included one at F563 (F563Y), which was previously found in a 2-hybrid screen to be somewhat defective for an interaction with AsiA³⁴. However, this and another screen did not reveal an importance for other region 4.1 residues that contact AsiA in the AsiA/region 4 structure ^{34; 35}. Thus, we constructed σ⁷⁰ mutants with alanine substitutions within region 4.1 that should either affect (σ⁷⁰551-552, 554-555A) or not affect (σ⁷⁰557, 560A) the interaction of AsiA with region 4 based on the AsiA/region 4 structure. For region 4.2, we made a σ⁷⁰ mutant that contained alanine substitutions at residues that directly interact with the –35 DNA element (σ⁷⁰584-588A). Although residues within this portion of region 4.2 interact with AsiA in the structure, a previous analysis of region 4.2 substitutions had not revealed an importance for this portion of 4.2 for AsiA binding ³⁶. Finally, we compared the functions of a fused σ⁷⁰/σ³⁸ protein that contains multiple changes within the C-terminal portion of σ⁷⁰ (σ⁷⁰/σ³⁸H5) with the function of a σ⁷⁰/σ³⁸ fusion (σ⁷⁰/σ³⁸ region 4) that contains multiple changes throughout region 4, including residues that are predicted to interact with AsiA, and with the function of the primary σ of pertussis (σ^pertussis) that despite multiple changes within region 4 from σ⁷⁰, retains all the predicted AsiA contact residues. The various σ⁷⁰ mutants were then tested in assays to determine the effects of the σ⁷⁰ mutations on AsiA inhibition, MotA/AsiA activation, and the interactions of σ⁷⁰ region 4 with AsiA, MotA, and the β-flap.

Substitutions within σ⁷⁰ region 4.1 that impair the interaction of σ⁷⁰ with AsiA do not affect MotA/AsiA activation at the T4 middle promoter P_uvsX

During a T4 infection, transcription from most T4 middle promoters, such as P_uvsX (Fig. 1B), is fully dependent on AsiA and MotA. However, in vitro wild type polymerase can use P_uvsX in the absence of MotA and AsiA, presumably because of the presence of a perfect σ⁷⁰ –10 element and partial matches to the σ⁷⁰ –35 element and to the UP element sequence (for binding by the α subunit of core ³⁷). Consequently, we used transcription from P_uvsX to monitor the effect of various σ⁷⁰ mutations on transcription by polymerase alone and with AsiA and/or MotA (Fig. 2). As has been seen previously ³⁸, polymerase containing wild type σ⁷⁰ yields a moderate level of P_uvsX transcription in vitro, which is completely inhibited by the presence of AsiA. Addition of both AsiA and MotA yields a high level of activated transcription. In addition, in single round transcription assays and under certain conditions, such as the ones used here, we have found about a 2-fold increase in transcription in the presence of MotA without AsiA (Makela and Hinton, manuscript in preparation).

Fig. 2.

Transcriptional activities of polymerases reconstituted with mutant σ proteins using P_uvsX in vitro in the presence of AsiA and/or MotA. The indicated σ (0.2 pmol) and AsiAHis₆ (3 pmol ) were incubated in 1.8 μl of protein buffer I for 10 min at 37° C. A solution containing 0.05 pmol of core in 1.1 μl of protein buffer II was then added, and the resulting solution was incubated for an additional 10 min at 37° C. Transcription was initiated by adding this solution to 2.1 μl of DNA buffer containing 0.01 pmol DNA. After 37° C for 20 sec, rifampicin (150 ng) was added to limit transcription to a single round. For each indicated σ, [the amount of P_uvsX RNA observed]/[the amount of P_usvX RNA observed with σ⁷⁰ polymerase in the absence of MotA and AsiA] is shown using polymerase alone, polymerase with MotA, polymerase with AsiA, and polymerase with AsiA and MotA. Values were determined from two or more transcriptions. Error is indicated by darker shaded area. In some cases where the level of transcription was low, 10-fold values (10 × values) are also shown. In these cases, zero is the bottom of the enclosed box.

Like wild type polymerase, the polymerases reconstituted with any of the σ proteins having substitutions within region 4.1 (σ⁷⁰551-552,554-555A, σ⁷⁰557,560A, σ⁷⁰555,557,560A, σ⁷⁰F563Y) were fully competent for MotA/AsiA activation at P_uvsX (Fig. 2). Thus, none of these mutations impaired σ appropriation at P_uvsX. In contrast, AsiA inhibition was not observed when using σ⁷⁰551-552, 554-555A (Fig. 2A) and as has been reported previously ³⁴, AsiA inhibition was impaired by σ⁷⁰F563Y (Fig. 2B). (Although the level of transcription with σ⁷⁰551-552, 554-555A was extremely low, the results were highly reproducible.)

The ability of the σ⁷⁰ mutants to interact with AsiA and MotA was investigated by two independent assays of protein-protein interactions. In the native gel assay, free σ was incubated with excess AsiA and/or MotA and the formation of protein-protein complexes was detected by a shift in σ migration. Previous work ^{39; 40} has shown that either AsiA or MotA alone forms a complex with wild type σ⁷⁰ and a new species is formed in the presence of all three, indicative of a tertiary complex (Fig. 3). In the E. coli 2-hybrid assay, the interaction of σ⁷⁰ region 4 with a bait protein, here AsiAK20A or the N-terminal half of MotA (MotA^NTD), which is known to contain the activation domain ²⁰, was monitored by an increase in the level of activity of a reporter gene, lacZ (Fig. 4A). AsiA with the substitution K20A was used in this assay because previous work has shown that AsiAK20A results in a higher level of β-galactosidase activity when tested with σ⁷⁰ region 4 than does wild type AsiA ³⁴.

Fig. 3.

σ/AsiA, σ/MotA, and AsiA/σ/MotA complexes in native protein gels. The indicated σ was incubated with excess His₆AsiA and/or MotA and then subjected to electrophoresis through native 6 % acrylamide gels as described in Materials and Methods. The positions of σ, the σ/AsiA complex, the σ/MotA complex, and the AsiA/σ/MotA complex are shown. The positions of free MotA and AsiA are also marked. Excess free MotA protein is the dark band that migrates near the top of the gel in the lanes with added MotA. The excess free AsiA is usually not seen because AsiA is a small protein (10 kDa) and stains poorly.

Fig. 4.

2-hybrid assay and the effect of σ⁷⁰ substitutions on the interaction of σ⁷⁰ region 4 with the AsiAK20A and MotA^NTD. (A) Cartoon shows the test promoter, which has a λ operator (O_R) located 62 bp upstream from the initiation site of the lac core promoter. The presence of the α-σ chimera protein, expressed by a pBRα-σ plasmid, and the cI-bait fusion protein, expressed by a pcI-bait plasmid, can result in an increase in lacZ transcription and subsequent β-galactosidase activity if there is an interaction between the σ present in the chimera and the bait protein. The addition of IPTG induces expression of both the α-σ chimera and the cI-bait fusion protein. B-E. Effect of σ⁷⁰ mutations on the interaction of σ⁷⁰ region 4 with AsiAK20A (left panels) and MotA^NTD (right panels). Cultures contained the pcI-AsiAK20A or pcI-MotA^NTD plasmid and either the control plasmid, pBRα that lacks σ⁷⁰ region 4, or as indicated, a pBRα-σ⁷⁰ plasmid that has σ⁷⁰ region 4 (σ⁷⁰ 4), σ⁷⁰ region 4.1 (σ⁷⁰ 4.1), σ⁷⁰ region 4.2 (σ⁷⁰ 4.2) with or without the indicated σ⁷⁰ substitutions. Graphs show β-galactosidase activity (in Miller units) versus the concentration of IPTG present in the cultures.

Interaction assays indicated that alanine substitutions at σ⁷⁰ residues 555, 557, and 560 had little effect on the interaction of σ⁷⁰ with AsiA or MotA. The pattern of complexes seen in the native gel with σ⁷⁰557,560A was very similar to that seen with wild type σ⁷⁰ (Fig. 3). 2-hybrid assays indicated that the σ⁷⁰ 555, 557, 560A substitutions had no effect on the interaction with MotA^NTD (Fig. 4B, right) and only a slight decrease in the interaction with AsiAK20A (Fig. 4B, left). These results are consistent with the fact either σ⁷⁰ 555, 557, 560A or σ⁷⁰ 557, 560A yielded transcription patterns that were very similar to that of wild type σ⁷⁰ (Fig. 2A).

Results obtained in the interaction assays with the other 4.1 mutants, σ⁷⁰ 551-552, 554-555A and σ⁷⁰ F563Y, were different. Previous results have shown that the σ⁷⁰ F563Y substitution renders σ⁷⁰ region 4 defective in its interaction with AsiAK20A in the 2-hybrid assay ³⁴. We found that the 551-552, 554-555A substitutions also impaired the interaction of σ⁷⁰ with AsiAK20A (Fig.4C, left). In contrast, these substitutions had a more modest effect on the σ⁷⁰ region 4/MotA^NTD interaction (Fig. 4C, right). In native gels, σ⁷⁰ 551-552, 554-555 and σ⁷⁰ F563Y were very slightly retarded when AsiA was added (Fig. 3). In the presence of MotA, a much slower migrating species was formed, consistent with the formation of a stable MotA/σ⁷⁰ complex. Addition of AsiA did not significantly change the position of this species, although a very slight shift (like that seen with AsiA alone) would not have been detected. We conclude that the 551-552, 554-555A and the F563Y substitutions result in a σ⁷⁰ that is impaired in its interaction with AsiA but not with MotA. Consequently, σ⁷⁰ mutant proteins having these substitutions are less susceptible to inhibition by AsiA because they are less able to interact with AsiA. However, the defect for AsiA binding does not impair MotA/AsiA activation. Thus, during transcription either the presence of MotA can improve the defective interaction of AsiA with σ⁷⁰ 551-552, 554-555A or σ⁷⁰ F563Y or an impaired interaction of AsiA with region 4.1 is still sufficient for AsiA to work as a co-activator.

We also used the two-hybrid assays to investigate the interactions of AsiAK20A and MotA^NTD with region 4.1 alone (Fig. 4D). We found that AsiAK20A interacts weakly with region 4.1 in the absence of region 4.2. Surprisingly, this interaction was not disrupted by the 551-552, 554-555A substitutions, suggesting that it might arise through different contacts than those used when AsiA interacts with all of region 4. In addition, we found that σ⁷⁰ region 4.1 alone was not sufficient for an interaction with MotA^NTD.

Alanine substitutions at residues 584-588 within σ⁷⁰ region 4.2 render σ⁷⁰ defective for an interaction with AsiA and for AsiA inhibition and MotA/AsiA activation at P_uvsX

The side chains of σ⁷⁰ 4.2 residues R584, E585, and R588 are thought to interact directly with specific base determinants in the –35 sequences of host promoter DNA ^{8; 35; 41; 42; 43}. Residues R584 and I587 were identified as contact residues for AsiA in the AsiA/σ⁷⁰ region 4 structure ³⁰ (Fig. 1A), suggesting that AsiA prevents the binding of σ⁷⁰ with the –35 element in part by directly blocking the σ⁷⁰ region 4/DNA interaction. However, a previous analysis, investigating how single alanine substitutions affect the binding AsiA to σ⁷⁰ region 4.2 peptides, had found that residues 584-588 were not important for the interaction of AsiA with region 4.2 ³⁶. To determine the importance of this portion of region 4.2 when present within all of region 4, we constructed σ⁷⁰ with alanine substitutions at residues 584-588. In the 2-hybrid assay, these mutations eliminated the interaction of AsiAK20A with σ⁷⁰ region 4 (Fig. 4C, left), indicating that one or more of residues 584-588 are indeed important for an interaction with AsiA. However, we also found that these 4.2 contacts are not sufficient because the interaction of σ⁷⁰ residues 569-613 (region 4.2 plus the C-terminal end of σ⁷⁰) with AsiA was low (Fig. 4E, left). Results with the native gel assay were difficult to interpret because σ⁷⁰584-588A migrated very differently from wild type σ⁷⁰ (Fig. 3). Nevertheless, the addition of AsiA did not significantly affect this migration, a result that was consistent with the lack of a detectable interaction between AsiA and σ⁷⁰ 584-588A in the 2-hybrid assay.

In contrast to the results with AsiAK20A, the σ⁷⁰584-588A did not impair the interaction of σ⁷⁰ with MotA^NTD. In 2-hybrid assays with all of σ⁷⁰ region 4 and MotA^NTD, the 584-588 alanine substitutions gave results that were similar to wild type region 4 (Fig. 4C, right). When using a portion of σ⁷⁰ containing residues 569-613 (region 4.2 to the end of σ⁷⁰) (Fig. 4E, right), the presence of the substitutions even increased its interaction with MotA^NTD. In native gels, the migration of σ⁷⁰584-588A was shifted in the presence of MotA without AsiA and the addition of AsiA did not change this migration (Fig. 3). The σ⁷⁰ 584-588A/MotA complex migrated differently from the complex formed with wild type σ⁷⁰ and MotA, again perhaps because the mutant σ alone migrated aberrantly.

In in vitro transcription assays, P_uvsX transcription by Eσ⁷⁰584-588A was not inhibited by AsiA alone or activated by MotA/AsiA (Fig. 2A). (The 2-fold increase seen with MotA/AsiA is not significant because it is also observed with MotA alone). We conclude that alanine substitutions at σ⁷⁰ residues 584-588 specifically impair the interaction of AsiA with σ⁷⁰ region 4, and this impairment then eliminates AsiA inhibition and MotA/AsiA activation at P_uvsX.

Mutation of the far C-terminal region of σ⁷⁰ significantly reduces MotA/AsiA activation

Previous work has indicated that the far C-terminal region of σ⁷⁰ is needed for an interaction between MotA^NTD and σ⁷⁰, ²⁰. Specifically, in the 2-hybrid assay, a σ⁷⁰/σ³⁸H5 mutation, in which the last 17 residues of σ⁷⁰ (Helix 5 (H5) in Fig. 1A) have been replaced with the corresponding residues of σ³⁸, eliminates the interaction of MotA^NTD with σ⁷⁰ region 4, but decreases the interaction with AsiA by only about 50% ²⁰. When we tested this mutant in transcription assays, we found that AsiA inhibition was normal and the modest increase in transcription observed with MotA alone was present, but MotA/AsiA activation was significantly reduced (Fig. 2B). This result supports the idea that during σ appropriation, the interaction of σ⁷⁰ H5 with MotA is crucial for activation. However, the increase in transcription observed with MotA alone as well as the low level of MotA/AsiA activation that was observed with σ⁷⁰/σ³⁸H5 also suggest that either a very poor interaction of σ³⁸ H5 with MotA is still somewhat productive for transcription or there is another interaction site for MotA, besides H5, that can promote transcription.

We also tested σ⁷⁰Δ571-613, which removes the far C-terminal region of σ⁷⁰ as well as all of region 4.2. Previous work with a similar deletion (σ⁷⁰Δ 565-613) had indicated that when only region 4.1 is present, AsiA alone significantly increases transcription from the T4 middle promoter P_rIIB2 ⁴⁴. Thus, we wondered whether a similar result would be observed with P_uvsX. However, in this case, transcription was unaffected by the addition of AsiA or MotA, and there was only a small, but reproducible increase in transcription when both MotA and AsiA were present (Fig. 2A). As was discussed before, in 2-hybrid assays the interaction of region 4.1 alone with MotA^NTD was nearly background and the interaction with AsiA was significantly lower than that seen with all of region 4 (Fig. 4D). In native gels, the migration of σ⁷⁰Δ571-613 was unaffected by the addition of MotA, AsiA, or AsiA and MotA (Fig. 3). These results are consistent with the idea that the deletion of regions 4.2 and H5 significantly impairs the interaction of σ⁷⁰ with either AsiA or MotA. However, the slight increase in transcription that is observed when both MotA and AsiA are present suggests that a weak interaction of AsiA with region 4.1 and an interaction of MotA with a region other than H5 can weakly promote transcription.

Replacement of σ⁷⁰ region 4 with region 4 of σ³⁸ eliminates interactions with AsiA and MotA while the primary σ of B. pertussis is competent for MotA/AsiA activation

As seen in Fig. 1A, region 4 of σ³⁸ differs from region 4 of σ⁷⁰ at many residues, including those that interact with AsiA in the AsiA/σ⁷⁰ structure and those that include the MotA site in H5. The addition of either AsiA alone or MotA and AsiA together had no significant effect on P_uvsX transcription when using Eσ⁷⁰/σ³⁸region 4 (Fig. 2B), indicating that σ³⁸ region 4 is not competent for AsiA inhibition or MotA/AsiA activation. Furthermore, the 2-fold increase in P_uvsX transcription seen with MotA alone was also eliminated when using Eσ⁷⁰/σ³⁸region 4. When combined with the results with σ⁷⁰/σ³⁸H5 and σ⁷⁰Δ571-613, these results suggest that another interaction site for MotA may lie within region 4.1.

The primary σ of B. pertussis also varies from σ⁷⁰ at many residues, including several in region 4 (Fig. 1). However, all of the residues that contact AsiA in the σ⁷⁰ region 4/AsiA structure are identical in the E. coli and B. pertussis primary σ proteins. We found that Eσ^pertussis was significantly inhibited by AsiA (Fig. 2B), which is consistent with known AsiA/σ⁷⁰ contacts. In contrast, the far C-terminal region (H5) of the pertussis σ differs at several residues from that of σ⁷⁰. Despite these differences, we found that Eσ^pertussis was significantly activated by MotA/AsiA (Fig. 2B). This result suggests that none of the differences between σ^pertussis and σ⁷⁰ in H5 are crucial for the σ⁷⁰/MotA interaction.

Effect of σ⁷⁰ mutations on the interaction of σ⁷⁰ region 4 with the β-flap

Because many residues within region 4 are thought to interact with the β-flap ^{10; 11; 12; 13} and AsiA is thought to prevent the σ⁷⁰/β-flap interaction ³⁴, we investigated the effect of two of the σ⁷⁰ region 4 mutants on the interaction of region 4 with the β-flap in the 2-hybrid assay. Previous work has shown that the interaction of region 4 with the β-flap can be monitored using this assay, provided that σ⁷⁰ region 4 contains the substitution D581G ¹³. D581G is thought to stabilize the H-T-H conformation of region 4.2 (H3-T-H4 in Fig. 1A) and does not affect the interaction of region 4 with the -35 DNA element ^{13; 45}. In addition, previous work has also indicated that region 4 of σ³⁸, which contains a glycine at the position corresponding to σ⁷⁰ D581, is fully competent for the interaction with the β-flap ¹³. We found here that the alanine substitutions at 551-552, 554-555 eliminated the interaction of σ⁷⁰ region 4 with the β-flap (Fig. 5A). This result is consistent with the very low level of transcription observed by polymerase containing the σ⁷⁰ 551-552, 554-555A mutant in the absence of AsiA and MotA (Fig. 2A). However, this result was surprising since none of the substituted residues have been implicated in an interaction with the β-flap (Fig.1A) ^{45; 11; 12}. In contrast, the interaction of region 4 with the β-flap was actually increased by the alanine substitutions at 584-588 (Fig. 5B). We conclude that the significant reduction in transcription observed with polymerase containing σ⁷⁰ 584-588A is not due to its inability to interact with the β-flap. Thus, it is likely to stem from the loss of the side chains at R584, E585, and R588, which are thought to interact directly with base determinants in the –35 DNA ^{8; 35; 42; 43}.

Fig. 5.

The σ⁷⁰ substitutions 551-552,554-555A (A) eliminate the interaction of the β-flap with σ⁷⁰ region 4 whereas substitutions 584-588A (B) do not impair the interaction of σ⁷⁰ region 4 with the β-flap. Cultures contained a pBRα-σ⁷⁰ plasmid, which had σ⁷⁰ region 4 (σ⁷⁰ 4) with the indicated substitutions, and either the control plasmid, pACλcI32, the pcI plasmid that lacks a fused bait protein, or the pcI-β-flap plasmid (β-flap). Graphs show β-galactosidase activity (in Miller units) versus the concentration of IPTG present in the cultures.

In addition to the interaction between σ⁷⁰ region 4 and the β-flap, contact between σ⁷⁰ region 4.1 residues 555, 562, and 565 and another region of core, a portion of β′ (residues 393 to 402), has been proposed ^{10; 46}. However, our 2-hybrid assays using pcI-rpoC (357-430) and pBRα-σ⁷⁰ or pBRα-σ⁷⁰D581G failed to detect any interaction (data not shown).

Discussion

Interface of σ⁷⁰ region 4 with AsiA and MotA

T4 middle promoters are hybrid promoters, which contain an excellent match to the σ⁷⁰ –10 element, but replace the σ⁷⁰ –35 element with the MotA box sequence centered at –30 (reviewed in ⁴⁷). The T4 MotA and AsiA proteins are needed to switch the sequence specificity of polymerase for the upstream promoter DNA (reviewed in ¹⁶). For host promoters, polymerase recognition of a promoter sequence involves contact with the DNA but also the correct positioning of the DNA-binding domains within polymerase. In E. coli, recognition of -10/-35 promoters requires interaction between residues in σ⁷⁰ region 2 and a -10 element and between residues in regions 4.1 and 4.2 and a -35 DNA element, respectively ^{1; 8; 9}. (Fig. 6A; highlighted in pink in Fig. 1A). Specific base determinants in the –35 DNA are contacted by 4.2 residues R584 ^{8; 35; 42}, E585 ^{8; 43}, and perhaps R588 ⁴¹, which are present in the second helix of H3-T-H4 (Fig. 1A). The interactions between region 4 and the β-flap are also needed to position region 4 properly relative to region 2, allowing simultaneous contact of both the –10 and –35 elements ^{10; 11; 12; 13; 48}. σ⁷⁰ residues, which are needed for this contact, have been located within region 4.1 and at the C-terminus of σ⁷⁰, including H5 ^{11; 12; 34; 45} (highlighted in blue in Fig. 1A) Our work suggests that residues 551-552, 554-555 are also important. Alanine substitutions at these positions eliminate the interaction of σ⁷⁰ with the β-flap in a 2-hybrid assay (Fig. 5A) and result in very low levels of transcription by polymerase alone (Fig. 2A). Thus, assuming that the interaction between the β-flap peptide and σ⁷⁰ region 4 in the 2-hybrid assay reflects the interaction in holoenzyme, these results argue that residues 551-555 are involved in the σ⁷⁰/β-flap interaction or they affect the local conformation. The wild type level of activated P_uvsX transcription observed with σ⁷⁰551-552, 554-555A in the presence of AsiA (Fig. 2A) and MotA eliminates the possibility that the mutations are deleterious simply because they result in a protein that is generally misfolded.

Fig. 6.

Interaction of σ⁷⁰ region 4 with -35 element and the β-flap versus its interaction with AsiA and MotA. Structures of (A) T. aquaticus σ region 4 (residues corresponding to E. coli σ⁷⁰ residues 546-612) with the –35 region of promoter DNA ⁸ (PDB: 1KU7) and of (B) E. coli σ⁷⁰ region 4 (residues 546-613) with AsiA (residues 2-90) ³⁰ (PDB: 1TLH). σ⁷⁰ region 4 is shown in gold, AsiA is shown in green, and the DNA is shown in magenta. Alpha helices H2, H3, H4, and H5 in σ⁷⁰ (see Fig. 1A) are labeled. Spacefill structures of σ⁷⁰ residues 551, 554, and 555 in region 4.1 and residues 584 and 587 in region 4.2, which are predicted to interact with AsiA in the structure and whose substitution impairs the σ⁷⁰/AsiA interaction, are shown in red in (B). The portion of σ⁷⁰ region 4 that interacts with the β-flap is indicated by the blue oval in (A). H5 of σ⁷⁰ that interacts with MotA is indicated by the black rectangle in (B). Note that the binding of AsiA to σ⁷⁰ region 4 converts H3-T-H4 of σ⁷⁰, which normally interacts with the –35 sequences of promoter DNA, into one continuous helix.

Previous work has indicated that in its interaction with AsiA, σ⁷⁰ uses a broad interface that includes multiple residues in regions 4.1 and 4.2 (reviewed in ¹⁶). The interface revealed by the σ⁷⁰ region 4/AsiA structure was quite surprising since it showed a previously unseen conformation of region 4, a dramatic reconfiguration that changes the DNA-binding H3-T-H4 to one continuous helix. Prior biochemical analyses that had determined specific AsiA residues needed for the AsiA/σ⁷⁰ interaction ^{31; 49} were in good agreement with the structure (30 reviewed in ¹⁶), providing support for the AsiA interface. The biochemical analyses here now provide support for several σ⁷⁰ residues implicated by the structure. Alanine substitutions at σ⁷⁰ residues that interact with AsiA in the structure [L551, R554 and E555 (σ⁷⁰ 551-552, 554-555A mutant) or R584 and I587 (σ⁷⁰ 584-588A mutant)] seriously impair AsiA/σ⁷⁰ binding (Fig. 4C) and AsiA inhibition (Fig. 2A). Substitution of F563, which is also predicted to interact with AsiA, also decreases the AsiA/region 4 interaction and AsiA inhibition (Fig. 2B and ³⁴). In contrast, σ⁷⁰ mutants with substitutions of residues that are not predicted to interact (the σ⁷⁰567, 560A and σ^pertussis mutants) behave like wild type σ⁷⁰ (Fig. 2A and B, Fig. 3). Previous work has also shown that a single alanine substitution at T552 is somewhat defective for the σ⁷⁰ region 4/AsiA interaction and for AsiA inhibition ³⁵. Although T552 is not a contact point in the structure, it is adjacent to the implicated contact residue, L551.

The work here also indicates that the interaction of AsiA with region 4 does not appear to be a simple addition of its contacts with regions 4.1 and 4.2. AsiA does not interact with σ⁷⁰ region 4.1 or region 4.2 alone in a yeast 2-hybrid assay ⁵⁰, and we have seen only a low level of interaction with either region alone in the E. coli 2-hybrid assay (Fig. 4D and E). In addition, we found that the effect of the 551-552, 554-555A mutations in the 2-hybrid with AsiA differed whether we used region 4.1 alone (Fig. 4D) or all of region 4 (Fig. 4C). These results might explain the differences between the AsiA contacts reported in the NMR structure of AsiA with all of region 4³⁰ and an NMR analysis performed using AsiA and separate σ⁷⁰ peptides for region 4.1 (residues 540 to 565) and region 4.2 (residues 570 to 599) ²⁹.

Presently there is not a structure detailing the molecular interface between MotA and σ⁷⁰. Previous biochemical evidence has identified H5 as a region of contact for MotA ²⁰. We show here that MotA/AsiA activation is significantly reduced when using Eσ⁷⁰/σ³⁸H5, but not when using Eσ^pertussis (Fig. 2B). These results suggest that L607, S609, F610, and/or L611, the H5 residues that are identical between the E. coli and the pertussis primary σ factors, may be contact residues for MotA. In addition, the weak transcription observed with AsiA and MotA even when using σ⁷⁰Δ571-613 (Fig. 2A), in which H5 is missing, or σ⁷⁰/σ³⁸H5 (Fig. 2B), in which σ⁷⁰ H5 has been replaced with σ³⁸ H5, suggests that it is region 4.1 that may contain additional contacts for MotA.

Model for AsiA/MotA activation at P_uvsX

Based on structural 30 and biochemical ^{20; 31} work, a model for how MotA and AsiA together activate transcription at P_uvsX has emerged. AsiA associates with polymerase by binding tightly to free σ⁷⁰, and then remaining stably bound after σ⁷⁰ binds to core ^{24; 51; 52; 53}. The dramatic conformational change of region 4 that is seen once AsiA is bound would then be incompatible with the interaction of σ⁷⁰ with the –35 DNA and with the β-flap. Thus, AsiA alone acts as an inhibitor of transcription from promoters that require an interaction with the –35 DNA. Because AsiA prevents the interactions of σ⁷⁰ with the β-flap and the –35 DNA, AsiA binding also makes σ⁷⁰ H5 and the MotA box, centered at –30, available for MotA. Thus, AsiA binding could facilitate MotA activation simply because of its inhibition function. However, our results with the σ⁷⁰551-552, 554-555A mutant suggest that eliminating σ⁷⁰ interactions with the β-flap and with the –35 DNA is not sufficient for MotA activation at P_uvsX. When the 551-552, 554-555A mutations were present, we did not detect an interaction between region 4 and the β-flap (Fig. 5A), and transcription was severely compromised (Fig. 2A). These results argue that when polymerase contains σ⁷⁰551-552, 554-555A, the σ⁷⁰/β-flap interaction is disrupted and as a consequence, the σ⁷⁰ region 4/-35 DNA interaction is also lost. However, despite these defects, AsiA was still required for MotA activation (Fig. 2A). Our results suggest that AsiA provides an additional function for MotA apart from disrupting the region 4/β-flap and region 4/DNA contacts. The AsiA-induced conformational change of H5, observed in the AsiA/σ⁷⁰ structure ³⁰, would be an additional feature of AsiA binding that could be needed to facilitate MotA interaction with H5 and indirectly the interaction of MotA with MotA box DNA.

Previous work has shown that in the absence of MotA, AsiA alone increases in vitro transcription from the T4 middle promoter P_rIIB2 when using a σ⁷⁰ that is truncated after region 4.1 at residue 565 ⁴⁴. This result suggested that MotA may function to remove the interaction of σ⁷⁰ H5 with the β-flap, allowing the interaction of AsiA with σ⁷⁰ region 4.1, which then activates transcription. Thus, when only region 4.1 is present, AsiA activates without MotA. However, we found that σ⁷⁰ truncated just after region 4.1 (Δ571-613) did not interact with AsiA in the native gel assay (Fig. 3) and did not yield AsiA activation of P_uvsX in the absence (or presence) of MotA (Fig. 2A). Thus, we do not detect activation by AsiA alone at P_uvsX like that seen at P_rIIB2. Further studies will be needed to test differences between various T4 middle promoters.

No other activators besides MotA are known to interact with σ⁷⁰ H5. While certain class II activators have been shown to interact with σ⁷⁰ region 4, they use the C-terminal half of H4 in region 4.2 ^{54; 55; 56}. However, H5 of σ³⁸ may be involved in regulating the activity of Eσ³⁸ at σ³⁸-dependent promoters needed to express genes required to cope with conditions of high salt ^{57; 58}. A thorough understanding of how MotA interacts with σ⁷⁰ H5 may shed light on how the C-terminal portion of σ factors can be used as a target for regulation.

Materials and Methods

DNA and Buffers

Plasmids were constructed using standard cloning procedures. Sequences of primers are available upon request. DNA sequencing (CBR-DNA Sequencing Facility of the University of Maryland, MWG Biotech, or UMDNJ (University of Medicine and Dentistry of New Jersey) DNA Core facility) throughout the cloned regions confirmed the presence of only the desired sequences.

pDKT90, which contains P_uvsX, has been described 59. pBRα-σ⁷⁰, ⁶⁰ and pBRα-σ⁷⁰ D581G ⁶¹ direct transcription of an α-σ chimera gene, composed of residues 1-248 of α fused in frame to residues 528-613 of σ⁷⁰, with or without the substitution at D581G. In each case, the gene is under the control of tandem promoters lpp and isopropyl-β-D-thiogalatoside (IPTG)-inducible lacUV5. pBRα ⁶⁰ encodes the N-terminal domain of wild type α (residues 1-248). pACλcI32 ⁶² encodes residues 1-236 of the bacteriophage λ cI protein under the control of the IPTG-inducible lacUV5 promoter. pACΔ-35λcI-AsiAK20A ³⁴ expresses a cI-AsiA fusion protein which contains the K20A substitution within the AsiA protein, and pcI-MotA^{NTD 20} expresses a cI-MotA^NTD protein in which the N-terminal domain of motA (encoding residues 1 to 97, followed by a non-coded serine) has been fused in frame to cI in pACλcI32. The K20A substitution in AsiA was used because this substitution results in a higher level of β-galactosidase activity in the 2-hybrid assay ³⁴. pACλcI-β-flap contains the rpoB gene encoding residues 858-946 of the β subunit of RNAP (the β-flap moiety) fused in frame to cI in pACλcI32 and has been described ¹³. pcI-rpoC (357-430), in which a portion of the rpoC gene encoding residues 357 to 430 of the β’ subunit of RNAP is fused in frame to cI in pACλcI32 was constructed by first obtaining a PCR product containing this region of rpoC using chromosomal DNA from E. coli XL1-Blue (Stratagene), a primer that annealed to rpoC DNA encoding residues 357 to 362 preceded by a NotI site, a primer that annealed to rpoC DNA encoding residues 430 to 425 preceded by a BglII site, and Pfu polymerase (Stratagene). After digestion with BglII and NotI, the fragment was ligated with pACλcI32 that had been digested with BglII and NotI, resulting in pcI-rpoC (357-430).

pETσ^fl was obtained by inserting the EcoRI/HindIII fragment containing σ⁷⁰ DNA from pσ^{fl 63} into pET21(+) (Novagen) that had been digested with EcoRI and HindIII. The resulting plasmid expresses His₆-tagged σ⁷⁰ protein, which has the sequence MRGSHHHHHHGSSGLVPRGSELGTRL fused to the start of σ⁷⁰. This σ⁷⁰ also contains the substitutions V36L and N149D, but we have not found any consequence of these substitutions in numerous assays (data not shown). To make pETσ^584-588A we used the Quikchange procedure (Stratagene) and primers that introduced codons for alanines at σ⁷⁰ residues 584-588 as well as introduced a PstI site. The presence of the desired substitutions in the resulting plasmids were screened by PstI digestion and DNA sequencing. The ClaI/HindIII fragment containing the 3′ end of σ⁷⁰ DNA was then obtained from a plasmid with the desired change, and this fragment was inserted into pETσ^fl that had been digested with ClaI and HindIII, resulting in pETσ584-588A. To construct pETσ551-552,554-555A, we first deleted the DNA coding for σ⁷⁰ residues 551-555 in pETσ^fl using the Quikchange procedure and primers that surrounded the DNA for residues 551-555, but lacked the 551-555 codons. Using the resulting deleted plasmid, we then used the Quikchange procedure with primers that surrounded the DNA for residues 551-555, but now had alanine codons for residues 551-555. After sequencing to identify a plasmid with the desired change, the XhoI fragment containing the 3′ end of the σ⁷⁰ gene was obtained and ligated to pETσ^fl that had been digested with XhoI, resulting in pETσ551-552,554-555A. pETσ⁷⁰/σ³⁸region 4 encodes a σ⁷⁰/σ³⁸ fusion protein in which σ⁷⁰ residues 1-529 are followed by σ³⁸ region 4 (residues 245-330). To construct this plasmid, we used the pBRα-σ³⁸; ⁶⁰, in which DNA encoding region 4 of σ³⁸ had been fused to DNA encoding the N-terminal domain of the α subunit of RNA polymerase. An XhoI site had also been placed at the α-σ junction in this plasmid. We performed PCR using pBRα-σ³⁸, a primer that annealed to the DNA at the α-σ³⁸ junction, and a primer that annealed to the DNA at the end of the σ³⁸ sequence and introduced a XhoI site. The resulting PCR product was digested with XhoI and ligated into pETσ^fl that had been digested with XhoI, resulting in pETσ⁷⁰/σ³⁸region 4. pETσ⁷⁰/σ³⁸H5 was constructed by ligating pETσ^fl that had been previously digested with ClaI and HindIII with the ClaI/HindIII fragment containing the σ³⁸ sequences from pBRα-σ⁷⁰/σ³⁸; ²⁰. pBRα-σ⁷⁰/σ³⁸ is identical to pBRα-σ⁷⁰ except that the σ⁷⁰ sequences encoding amino acids 597-613 have been replaced with the comparable σ³⁸ sequences (encoding amino acids 312-330). pETσ^Δ571-613 was constructed by ligating pETσ^fl that had been digested with ClaI and ScaI with the ClaI/ScaI fragment containing the 3′ end of σ⁷⁰ DNA from pσ^1-570, ²⁰.

To construct pET15b-557A560A and pET15b-555A557A560A, we used a plasmid pET15b-σ⁷⁰, in which the rpoD gene is cloned between NdeI and EcoRI sites. We performed two-step PCR site-directed mutagenesis and obtained PCR products containing the corresponding mutations and BamHI and EcoRI sites at the ends of the each PCR fragment. As the pET15b-σ⁷⁰ contains unique BamHI site in the rpoD gene, the plasmids pET15b-557A560A and pET15b-555A557A560A were obtained by inserting the corresponding BamHI/EcoRI fragments into the pET15b-σ⁷⁰ that had been digested with BamHI and EcoRI. These plasmids produce N-terminally His₆-tagged σ⁷⁰ with the indicated mutations.

To construct pBRα-σ^4.1 we obtained a PCR product with pETσ^fl and a primer that annealed to DNA encoding σ⁷⁰ residues 528-534 preceded by an XhoI site and a primer that annealed to DNA encoding 562-568 followed by a BamHI site. After digestion of the product with BamHI and XhoI, the fragment was ligated with pBRα-σ⁷⁰ that had also been digested with BamHI and XhoI. The resulting plasmid produces an α-σ^4.1 fusion protein, in which α residues 1-248 are followed by σ⁷⁰ residues 528-568. pBRα-σ^4.2 produces an α-σ^4.2 fusion protein, in which α residues 1-248 are followed by the protein sequence LEL (brought in by the linker DNA) and then σ⁷⁰ residues 569-613. It was constructed like pBRα-σ^4.1 except that the primers annealed to σ⁷⁰ residues 569-574 and 608-613. pBRα-σ^4.2584-588A, pBRα-σ^4.1551-552, 554-555A, and pBRα-σ⁴⁷⁰555,557,560A were constructed following similar procedures except that the PCR templates were pETσ⁷⁰584-588A, pETσ⁷⁰551-552, 554-555A, and pET15b-555A557A560A, respectively.

Protein buffer I contained 47 mM Tris-Cl (pH 7.9), 83 mM NaCl, 45% glycerol, 0.9 mM EDTA, 0.09 mM DTT, 0.009% Triton X-100, and 22 mM immidazole. Protein buffer II contained 5.2 mM Tris-Cl (pH 7.9), 72 mM Tris-acetate (pH 7.9), 26 mM NaCl, 5.2% glycerol, 0.2 mM EDTA, 0.3 mM DTT, 270 mM potassium glutamate, 7.2 mM magnesium acetate, and 180 μg/ml BSA. DNA buffer contained 53 mM Tris-acetate (pH 7.9), 6 mM NaCl, 0.2% glycerol, 0.6 mM EDTA, 0.1 mM DTT, 200 mM potassium acetate, 5.3 mM magnesium acetate, 130 μg/ml BSA, 0.48 mM each ATP, GTP, and CTP, and 24 μM [α-³²P]UTP (6 × 10⁴ dpm/pmol).

Proteins

The purification of wild type σ⁷⁰ (not His₆-tagged), which was used for the native gel assays, has been described ³⁹. Purification of the His₆-tagged proteins, σ⁷⁰, σ⁷⁰551-552,554-555A, σ⁷⁰/σ³⁸H5, and σ⁷⁰Δ571-613, was done using BL21(DE3)/pLysE ⁶⁴ cultures containing the desired pETσ plasmid as described ⁶⁵ by denaturation of inclusion bodies containing the protein, Ni⁺⁺ resin affinity chromatography under denaturing conditions, and then a slow renaturation of the protein. His₆-tagged σ⁷⁰/σ³⁸region 4 was purified by a similar procedure except that after sonication of the cells, this protein was found in the 17,400 × g supernatant rather than in inclusion bodies and consequently, it was subjected to Ni⁺⁺ resin chromatography under native conditions. 584-588A was purified by both procedures. Transcription assays His₆-tagged σ⁷⁰ indicated that the His₆-tagged proteins σ⁷⁰551-552, 554-555A, σ⁷⁰584-588A, σ⁷⁰/σ³⁸region 4, σ⁷⁰Δ571-613, and the σ⁷⁰ saturated core at a ratio of ~2:1 (data not shown). σ⁷⁰F563Y, σ^pertussis, and the σ⁷⁰ used as the control for the σ⁷⁰F563Y and σ^pertussis proteins (Fig. 5B) were gifts of Phil Boucher and Scott Stibitz (FDA). E. coli RNA polymerase core was purchased from Epicentre Technologies. MotA was purified as described ³⁸. The purifications of N-terminally His₆-tagged AsiA, His₆AsiA 31, and C-terminally His₆-tagged AsiA, AsiAHis₆ ⁴⁰ have been described. We have not observed any significant difference in activity between wild type AsiA, AsiAHis₆, and His₆AsiA.

To isolate the His₆-tagged proteins, σ⁷⁰555, 557, 560A and σ⁷⁰557, 560A, E. coli BL21(DE3) containing the appropriate plasmid was grown in LB broth with 100 μg/ml ampicillin at 37° C to an OD₆₀₀ of 0.8-1.0, and synthesis of the mutant σ⁷⁰ was induced by the addition of IPTG to 0.5 mM. After 6 more hours of growth, cells were harvested by centrifugation, resuspended in buffer A (20 mM Tris-HCl, pH 8.0; 500 mM NaCl; 5% glycerol; 2 mM imidazole, pH 8.0) and lysed by sonication. After centrifugation, the cleared lysate was loaded onto a 5 ml chelating Hi-trap Sepharose column attached to syringe. The column was then washed with buffer A containing 20 mm imidazole, and the proteins were eluted with 100 mm imidazole in the same buffer. The proteins were concentrated to 1–2 mg/ml by dialyzing against buffer B that contained 20 mM Tris-HCl, pH 8.0; 100 mM NaCl, 1 mM β–mercaptoethanol, 50 % glycerol, and stored at −20°C.

Two-hybrid assays

β-galactosidase assays were performed as described previously ³¹ using cultures of E. coli KS1 ⁶⁶ containing the indicated plasmids except that cultures were grown for 3 hr in LB media supplemented with the appropriate antibiotics at the following concentrations: carbenicillin (50 ug/ml), chloramphenicol (25 ug/ml), and kanamycin (50 ug/ml). β-galactosidase values represent the averages of duplicate assays. Error bars are shown in Fig. 4 and 5, but in some cases they are not seen because they are smaller than the symbols.

Native Protein Gels

Protein-protein complexes were separated from free proteins by gel electrophoresis on native, 6% polyacrylamide gels as described ³⁹. Reactions were composed of 24 pmol of His₆AsiA in 2 μl of a buffer containing 20 mM Tris-Cl (pH 7.9), 500 mM NaCl, 5% glycerol, and 100 mM immidazole); 4.6 pmol of σ⁷⁰ in 2 μl of a buffer containing 50 mM Tris-Cl (pH 8.0), 50 mM NaCl, 50% glycerol, 1 mM EDTA, 0.1 mM DTT, and 0.01% Triton X-100; and/or 17 pmol of MotA in 1 μl of a buffer containing 20 mM Tris-Cl (pH 7.9), 270 mM NaCl, 10% glycerol, 1 mM EDTA, and 1 mM 2-mercaptoethanol. Reactions were incubated at 37° C for 10 min before electrophoresis. After electrophoresis, proteins were detected using SilverXpress^® (Invitrogen).

In vitro transcriptions

Transcription reactions were assembled as indicated in the figure legends. Upon addition of rNTPs, reactions were incubated at 37° C for 20 sec, rifampicin (0.5 μl of 300 ng/μl) was added, and then the reactions were incubated for an additional 7 min before being collected on dry ice. Gel load solution (1 × TBE, 7 M urea, 0.1% bromphenol blue, 0.1% xylene cyanol FF) was added at a volume of 5 times that of the reaction aliquots, and the solution was heated at 95° C for two minutes before electrophoresis on 4% polyacrylamide, 7 M urea denaturing gels run in 1/2 × TBE. After autoradiography, films were scanned using a Powerlook 2100XL densitometer and quantification was performed using Quantity One software from Bio-Rad, Inc.