The interaction between σ⁷⁰ and the β-flap of Escherichia coli RNA polymerase inhibits extension of nascent RNA during early elongation

Abstract

The σ-subunit of bacterial RNA polymerase (RNAP) is required for promoter-specific transcription initiation. This function depends on specific intersubunit interactions that occur when σ associates with the RNAP core enzyme to form RNAP holoenzyme. Among these interactions, that between conserved region 4 of σ and the flap domain of the RNAP β-subunit (β-flap) is critical for recognition of the major class of bacterial promoters. Here, we describe the isolation of amino acid substitutions in region 4 of Escherichia coli σ⁷⁰ that have specific effects on the σ⁷⁰ region 4/β-flap interaction, either weakening or strengthening it. Using these σ⁷⁰ mutants, we demonstrate that the σ region 4/β-flap interaction also can affect events occurring downstream of transcription initiation during early elongation. Specifically, our results provide support for a structure-based proposal that, when bound to the β-flap, σ region 4 presents a barrier to the extension of the nascent RNA as it emerges from the RNA exit channel. Our findings support the view that the transition from initiation to elongation involves a staged disruption of σ-core interactions.

The bacterial RNA polymerase (RNAP) holoenzyme consists of a catalytic core enzyme (α₂ββ′ω) complexed with a σ factor. σ factors confer on the holoenzyme the ability to initiate promoter-specific transcription (1). The primary σ factor in Escherichia coli is σ⁷⁰, and a typical σ⁷⁰-dependent promoter bears two conserved sequence elements, the –10 and the –35 hexamers, that are separated by a spacer of ≈17 bp (1). All primary σ factors share four regions of conserved sequence (regions 1–4) (2). Regions 2, 3, and 4 contain DNA-binding domains responsible for recognition of the promoter –10 element, extended –10 element (3), and –35 element, respectively (1, 4). Regions 3 and 4 are separated by a flexible linker (region 3.2) (5, 6). σ⁷⁰ ordinarily recognizes promoter sequences only in the context of the holoenzyme, formation of which depends critically on a high-affinity interaction between σ region 2 and a coiled-coil motif in the β′-subunit (7, 8). A weaker interaction between σ region 4 and the flap domain of the β-subunit (β-flap) is not required for holoenzyme formation, but is required to position σ region 4 for sequence-specific interaction with the promoter –35 element (9).

During the transition from initiation to elongation, the interaction between σ and the remainder of the elongation complex is weakened (10–15). Structural evidence has led to the proposal that this is due, at least in part, to sequential displacement of portions of σ from the core enzyme by the newly synthesized RNA (5, 6, 16). Specifically, structures of RNAP holoenzyme (5, 6, 16) and structural models of the RNAP promoter open complex (16, 17) indicate that two regions of σ lie along the predicted path of the nascent RNA: (i) region 3.2, which is positioned within the RNA exit channel; and (ii) region 4, which, by virtue of its interaction with the β-flap, is positioned immediately adjacent to the end of the RNA exit channel. Thus, it has been proposed that, during early elongation, the nascent RNA (i) must displace σ region 3.2 from the RNA exit channel when the nascent RNA enters the channel (5, 16) [at a length of ≈10–11 nt (18)], and (ii) must displace σ region 4 from the β-flap when the nascent RNA emerges from the RNA exit channel (6) [at a length of ≈15–16 nt (19, 20)]. Analysis of early elongation complexes has provided evidence for displacement of σ region 3.2 from the RNA exit channel (21, 22), and possibly, for destabilization of the interaction between σ region 4 and the β-flap (23). However, evidence for the proposed RNA-mediated displacement of σ region 4 from the β-flap has not been reported.

Here, we describe the isolation and characterization of amino acid substitutions in region 4 of σ⁷⁰ that alter the strength of the σ⁷⁰ region 4/β-flap interaction. Using these substitutions, we show that the interaction between σ⁷⁰ region 4 and the β-flap can inhibit extension of the nascent RNA when it reaches a length of 16 nt. The results provide evidence for the proposed requirement to displace σ region 4 from the β-flap when the nascent RNA emerges from the RNA exit channel, at a length of ≈15–16 nt. As such, the results provide support for the view that the transition from initiation to elongation involves a staged disruption of σ -core interactions (4–6, 16, 18, 24).

Materials and Methods

Expression Vectors and Strains. Plasmid pACλcI-β-flap encodes residues 1–236 of the bacteriophage λ cI protein fused to residues 858–946 of β under the control of the isopropyl-β-d-thiogalactoside (IPTG)-inducible lacUV5 promoter (9). Plasmids pBRα-σ⁷⁰ and pBRα-σ⁷⁰ D581G encode residues 1–248 of α fused to residues 528–613 of σ⁷⁰ (with or without substitution D581G) under the control of tandem lpp and IPTG-inducible lacUV5 promoters (25, 26). Plasmid pBRα encodes wild-type α (27).

Strains KS1 (25) and BN317 (26) contain test promoters placO_R2–62 and placCons–35C, respectively, linked to lacZ present in single copy on the chromosome (KS1) or an F′ episome (BN317).

β-Gal Assays. β-gal assays were performed as described (28), in duplicate on separate occasions. Values are the averages from one experiment; duplicate measurements differed by <5%.

Proteins. His-tagged versions of wild-type and mutant σ⁷⁰ were purified as described (29) after overproduction from plasmid pLHN12-His. The E. coli RNAP core used in the assays shown in Figs. 4 and 5 was obtained from Epicentre Technologies (Madison, WI), and holoenzyme was made by incubation with a fivefold excess of σ⁷⁰. Plasmid pIA325 (30) coexpressing E. coli RNAP core subunits α, β′ and His₆-tagged βΔ900–909 was used to purify RNAP lacking the flap-tip helix (Δflap RNAP core) as described (31), except that a heparin purification step was omitted.

Fig. 4.

Effects of substitutions in σ⁷⁰ region 4 on transcription in vitro. Single-round in vitro transcription assays were performed as described in Supporting Methods by using RNAP holoenzyme reconstituted with either wild-type σ⁷⁰ or the indicated σ⁷⁰ mutant and linear DNA templates bearing one of the three indicated promoters (Left) or galP1/cons (Right). (Left) The run-off transcription products (123, 120, and 100 nt for T7A2, lacUV5, and galP1/cons, respectively) that were internally labeled by using α-[³²P]UTP. (Right) The abortive transcripts and the run-off transcripts (RO) that were end-labeled with γ-[³²P]ATP.

Fig. 5.

Effects of substitutions in σ⁷⁰ region 4 on early elongation pausing at λP_R′. (A) A linear DNA template carrying λP_R′ was used to perform single-round in vitro transcription assays as described in Supporting Methods with RNAP holoenzyme reconstituted with either wild-type σ⁷⁰ or the indicated σ⁷⁰ mutant. Aliquots of each transcription reaction were removed and stopped at the indicated times after the initiation of transcription. The transcripts were end-labeled with γ-[³²P]ATP. The 16- and 17-nt RNA species (+16 and +17) and the 194-nt terminated transcript (T) are indicated. A complete gel image is shown in Fig. 8, which is published as supporting information on the PNAS web site. (B) Shown is the ratio of paused elongation complexes containing a 17-nt transcript to paused elongation complexes containing a 16-nt transcript (+17/+16 ratio) for each holoenzyme species. These values are the means of three independent experiments including that shown in A and were calculated by using the 1-min time point.

Fluorescence-Detected Electrophoretic Mobility-Shift Competition Experiments. Reaction mixtures contained (40 μl): 200 nM fluorescein-labeled σ derivative (σ_F; labeled at position 366; prepared as in ref. 32), 0–600 nM unlabeled competitor sigma (σ_X; prepared as in ref. 29), and 100 nM wild-type RNAP core (prepared as in ref. 32) or 100 nM Δflap RNAP core (see above), 40 mM Tris·HCl (pH 8.0), 100 mM NaCl, 10 mM MgCl₂, 1 mM DTT, 0.2% Tween 20, and 5% glycerol. Reactions were initiated by addition of RNAP core and were allowed to proceed 12 h at 24°C. (Control reactions initiated by addition of σ_X yield equivalent results, indicating that the system reaches equilibrium.) Reaction mixtures were applied to 5% polyacrylamide slab gels, and electrophoresed in 90 mM Tris-borate, pH 8.0/0.2 mM EDTA (at 100 V for 90 min at room temperature). Fluorescence emission intensities were quantified by using an x/y fluorescence scanner (FluorImager 595, Molecular Dynamics). The fractional occupancy of RNAP core with σ_F (Θ) was calculated as described in Supporting Methods, which is published as supporting information on the PNAS web site.

Results

Substitutions in σ⁷⁰ Region 4 That Weaken or Strengthen Its Interaction with the β-Flap. To investigate how the interaction between σ region 4 and the β-flap affects events that occur during the transition from initiation to elongation, we sought to identify amino acid substitutions in region 4 of σ⁷⁰ that either weakened or strengthened the σ⁷⁰ region 4/β-flap interaction. To accomplish this goal, we took advantage of a bacterial two-hybrid assay (27, 28) that enabled us to detect the interaction between a fragment of σ encompassing region 4 and a fragment of β encompassing the flap domain (9). In this two-hybrid assay, contact between a protein domain fused to a subunit of RNAP and a partner domain fused to a DNA-binding protein activates transcription from a test promoter bearing a recognition site for the DNA-binding protein. In this case, region 4 of σ⁷⁰ (residues 528–613) was fused to the α-subunit N-terminal domain (NTD; replacing the α-subunit C-terminal domain), and the β-flap (residues 858–946) was fused to the cI protein of bacteriophage λ (Fig. 1A). Thus, we showed previously that the λcI-β-flap fusion protein activates transcription from test promoter placO_R2–62 specifically in the presence of the α-σ⁷⁰ chimera (9).

Fig. 1.

σ⁷⁰ region 4 substitutions that weaken its interaction with the β-flap. (A) Bacterial two-hybrid assay. Cartoon depiction of how the interaction between σ⁷⁰ region 4 (fused to the α-NTD) and the β-flap (fused to the bacteriophage λ cI protein) activates transcription from test promoter placO_R2–62, which bears the λ operator O_R2 centered 62 bp upstream of the start site of the lac core promoter. (B) Effects of substitutions in the σ moiety of the α-σ₇₀ D581G chimera on transcription from placO_R2–62 in the presence of the λcI-β-flap fusion protein. KS1 cells (see Materials and Methods) harboring compatible plasmids directing the synthesis of the indicated proteins were grown in the presence of different concentrations of IPTG and assayed for β-gal activity. (C) Schematic of test promoter placCons–35C. This promoter bears a consensus –35 element (TTGACA) upstream of the core promoter elements that serves as a binding site for the tethered σ⁷⁰ region 4 moiety. (D) Effects of substitutions in the σ moiety of the α-σ⁷⁰ D581G chimera on transcription from placCons–35C. BN317 cells (see Materials and Methods) harboring plasmids directing the synthesis of the indicated proteins were grown in the presence of different concentrations of IPTG and assayed for β-gal activity.

We first sought to identify amino acid substitutions in σ⁷⁰ region 4 that weakened its interaction with the β-flap. To accomplish this goal, we introduced random mutations into the gene fragment encoding the σ moiety of the α-σ⁷⁰ chimera and screened for mutations that specifically abolished the stimulatory effect of the λcI-β-flap fusion protein on transcription from test promoter placO_R2–62 (see Supporting Methods). To identify mutations causing effects that were specific to the σ⁷⁰ region 4/β-flap interaction, we took advantage of a genetic assay that we developed previously to detect the ability of σ region 4 to bind a –35 element (25, 26). In this assay, interaction between the σ moiety of the α-σ⁷⁰ chimera and an ectopic –35 element positioned upstream of the core promoter elements activates transcription from test promoter placCons–35C (26) (Fig. 1C). To facilitate the genetic screen, we also took advantage of a previously isolated amino acid substitution, D581G (26), that evidently stabilizes the folded structure of the tethered σ⁷⁰ region 4 moiety and thus facilitates detection of the interaction between σ⁷⁰ region 4 and both the β-flap and the ectopic –35 element (9, 26). We therefore screened for amino acid substitutions that disrupted the interaction between the σ moiety of the α-σ⁷⁰ D581G chimera and the β-flap (by using the assay depicted in Fig. 1 A), but did not disrupt the ability of the σ moiety to bind to a –35 element (by using the assay depicted in Fig. 1C).

Using these criteria, we identified several substitutions in the σ moiety of the α-σ⁷⁰ D581G chimera that specifically disrupted the σ⁷⁰ region 4/β-flap interaction (Fig. 1B and data not shown). For the purposes of performing functional studies, we selected two for further study. One substitution (L607P) completely abolished and the other substitution (R541C) nearly abolished the stimulatory effect of the λcI-β-flap fusion protein on transcription from placO_R2–62 (Fig. 1B), but neither inhibited the ability of the tethered σ moiety to bind to the ectopic –35 element of placCons–35C (Fig. 1D). We also combined the two substitutions and found that the effects of the double substitution were similar to that of single-substitution L607P (Fig. 1 B and D).

Next, we sought to identify amino acid substitutions in σ⁷⁰ region 4 that strengthened its interaction with the β-flap. Because the interaction between σ⁷⁰ region 4 (without the D581G substitution) and the β-flap is just above the threshold of detection in our two-hybrid assay (9), we used the α-σ⁷⁰ chimera containing wild-type σ⁷⁰ region 4 for this screen. After introducing random mutations into the gene fragment encoding the σ moiety of the α-σ⁷⁰ chimera, we screened for substitutions that increased the stimulatory effect of the λcI-β-flap fusion protein on transcription from placO_R2–62. As anticipated, we identified substitution D581G; among others that we identified (see Supporting Methods) was substitution T544M. Fig. 2A shows that substitutions D581G and T544M increased transcription from placO_R2–62 specifically in the presence of the λcI-β-flap fusion protein by factors of ≈3.5 and ≈2.5, respectively (compared with the level measured with the wild-type α-σ⁷⁰ chimera).

Fig. 2.

Substitutions in σ⁷⁰ region 4 that strengthen its interaction with the β-flap. (A) Effects of substitutions in the σ moiety of the α-σ⁷⁰ chimera on transcription from placO_R2–62 (depicted in Fig. 1 A) in the presence of the λcI-β-flap chimera. KS1 cells harboring compatible plasmids directing the synthesis of the indicated proteins were grown in the presence of different concentrations of IPTG and assayed for β-gal activity. (B) Effects of substitutions in the σ moiety of the α-σ⁷⁰ D581G chimera on transcription from placCons–35C (depicted in Fig. 1C). BN317 cells harboring plasmids directing the synthesis of the indicated proteins were grown in the presence of different concentrations of IPTG and assayed for β-gal activity.

An alignment of σ factors (2) revealed that in addition to threonine and methionine, the hydrophobic residues isoleucine, leucine, and valine also occur at the position corresponding to T544 of σ⁷⁰. We therefore tested the effect of introducing each of these hydrophobic residues (I, L, or V) at position 544 of the σ moiety of the α-σ⁷⁰ chimera. Substitution T544L had the same effect as substitution T544M on the interaction between σ⁷⁰ region 4 and the β-flap (data not shown), whereas substitutions T544I and T544V had stronger effects, increasing transcription from placO_R2–62 in the presence of the λcI-β-flap fusion protein by a factor of ≈8 (Fig. 2 A and data not shown). Finally, when introduced into the σ moiety of the α-σ⁷⁰ chimera in combination, substitutions D581G and T544I increased transcription from placO_R2–62 in the presence of the λcI-β-flap fusion protein by a factor of ≈40 (Fig. 2 A).

As a control, we tested the effects of the substitutions at σ⁷⁰ positions 581 and 544 on transcription from placCons–35C. As shown previously (26), substitution D581G in the tethered σ moiety permits detection of a protein/DNA interaction between the σ moiety and the ectopic –35 element of placCons–35C (Fig. 2B). In contrast, substitutions T544M and T544I did not have this effect (data not shown), and slightly reduced the ability of the tethered σ⁷⁰ D581G moiety to interact with the ectopic –35 element (Fig. 2B). Taken together, these results suggest that substitutions T544I and T544M specifically increase the strength of the σ⁷⁰ region 4/β-flap interaction.

Location of Amino Acid Substitutions in the Context of a Bacterial Holoenzyme Structure. Consistent with our observation that substitutions at positions R541, T544, and L607 specifically altered the σ⁷⁰ region 4/β-flap interaction, examination of the structure of a bacterial holoenzyme (6) reveals that the residues corresponding to σ⁷⁰ residues R541, T544, and L607 lie at the σ region 4/β-flap interface (Fig. 3), positioned for contact with residues of the β-flap-tip helix, the critical determinant for the binding of σ⁷⁰ region 4 (33). On the other hand, the residue corresponding to σ⁷⁰ residue D581 is not located at the σ region 4/β-flap interface (Fig. 3), which is consistent with our observation that effects of substitution D581G in σ⁷⁰ are not limited to the interaction of σ⁷⁰ region 4 with the β-flap (26).

Fig. 3.

Location of substitutions that alter the interaction between σ⁷⁰ region 4 and the β-flap in the Thermus thermophilus holoenzyme crystal structure (6). The T. thermophilus primary σ-subunit (σ^A) is orange, the β-subunit is cyan, and the β′-subunit is pink. Shown is the interface between σ region 4 (σ₄) and the β-flap-tip helix (cyan worm). The locations of residues in σ^A (351, 354, 391, and 418) that correspond to positions R541, T544, D581, and L607 of σ⁷⁰, respectively, are red (substitutions that weaken the interaction) or blue (substitutions that strengthen the interaction).

Effects of σ⁷⁰ Substitutions on Binding to RNAP Core in Vitro. We wished to determine whether amino acid substitutions identified in the two-hybrid assay would affect the σ⁷⁰/β-flap interaction in the context of RNAP holoenzyme. Accordingly, we performed fluorescence-detected EMSAs to quantify the effects of these substitutions on the σ⁷⁰/core interaction. We performed assays, in parallel, by using wild-type RNAP core and an RNAP core derivative lacking the β-flap-tip helix (30) (Δflap RNAP), to determine the specific effects of the substitutions on the σ⁷⁰/β-flap interaction (Fig. 7, which is published as supporting information on the PNAS web site). Thus, we compared the abilities of unlabeled wild-type σ⁷⁰ and mutant σ⁷⁰ derivatives to displace fluorescein-labeled wild-type σ⁷⁰ from either wild-type RNAP core or Δflap RNAP core, determining a set of relative equilibrium-binding constants (relative K_b values) for each RNAP core enzyme (Table 1, second and third columns). We then determined the ratios of the relative K_b values, thereby extracting the β-flap-dependent effects of the amino acid substitutions in σ⁷⁰ region 4 (Table 1, last column). Assuming that the only relevant difference between the two core enzymes is the inability of the Δflap RNAP core to bind σ⁷⁰ region 4, the results indicate that substitutions R541C, L607P, and the combination R541C/L607P weaken the σ⁷⁰/β-flap interaction in the context of RNAP holoenzyme, and that substitutions T544M, T544I, and the combination T544I/D581G strengthen the σ⁷⁰/β-flap interaction in the context of RNAP holoenzyme. In addition, the results establish a correlation between the relative effects of these substitutions in the two-hybrid assay in vivo and their relative effects in the competition assay in vitro.

Table 1. Relative equilibrium-binding constants.

	Wild-type RNAP	Δflap RNAP	Ratio
σ70 Derivative	Relative Kb	Relative Kb	Flap-specific relative Kb
R541C/L607P	0.46 ± 0.03	1.2 ± 0.2	0.38
L607P	0.52 ± 0.07	0.96 ± 0.09	0.54
R541C	1.0 ± 0.2	1.8 ± 0.2	0.56
Wild-type	1.0	1.0	1.0
T544M	2.8 ± 0.3	1.1 ± 0.1	2.5
T544I	5.9 ± 0.4	1.1 ± 0.1	5.4
T544I/D581G	12 ± 4	1.0 ± 0.1	12

Effects on Promoter Recognition. Next, we tested whether substitutions R541C and L607P would cause defects in the recognition of –10/–35 promoters. By using reconstituted holoenzymes (Eσ⁷⁰ WT, Eσ⁷⁰ R541C, etc.), we performed single-round run-off transcription assays with DNA templates bearing two different –10/–35 promoters: T7A2 and lacUV5. We found that, although Eσ⁷⁰ R541C and Eσ⁷⁰ L607P manifested either no defects (R541C) or modest defects (L607P) in the production of run-off transcripts from the –10/–35 promoters, Eσ⁷⁰ R541C/L607P produced only ≈40% and 10% as many transcripts as Eσ⁷⁰ WT from the T7A2 and lacUV5 promoters, respectively (Fig. 4). These results are consistent with the expectation that disrupting the σ⁷⁰ region 4/β-flap interaction in the context of the holoenzyme would cause a defect in the recognition of –10/–35 promoters (9, 33).

We also examined the ability of the mutant holoenzymes to transcribe a DNA template bearing a promoter of the extended –10 class (galP1/cons); extended –10 promoters are defined by the presence of a TG dinucleotide 1 bp upstream of the –10 hexamer, and typically lack a recognizable –35 element. As shown in Fig. 4, the mutant holoenzymes manifested no defects in the production of run-off transcripts from galP1/cons. However, further inspection of the transcription products synthesized under the control of galP1/cons revealed a difference between the mutant holoenzymes and Eσ⁷⁰ WT. As shown in Fig. 4 Right, several short RNA products (likely the result of abortive initiation, which occurs before promoter escape) are evident in the reaction performed with Eσ⁷⁰ WT; one of these, identified as the 9-mer based on analysis with limiting sets of nucleotides (data not shown), is particularly prominent. Strikingly, this RNA product is selectively depleted in reactions performed with Eσ⁷⁰ L607P (Fig. 4) or with Eσ⁷⁰ R541C/L607P (data not shown). These findings suggest that the weakened σ⁷⁰ region 4/β-flap interaction facilitates promoter escape at galP1/cons, raising the possibility that this effect (which results in a relative increase in the amount of run-off transcript) may mask a promoter-recognition defect at galP1/cons. To evaluate this possibility, we performed DNase I footprinting assays to compare the abilities of Eσ⁷⁰ WT and Eσ⁷⁰ L607P to bind galP1/cons; the mutant holoenzyme exhibited a moderate promoter-binding defect in this experiment (data not shown). These results suggest that the σ⁷⁰ region 4/β-flap interaction can contribute to promoter binding, even in the absence of an identifiable –35 element (see Discussion).

Effects on Early Elongation Pausing. The interaction between σ region 4 and the β-flap positions region 4 immediately adjacent to the end of the RNA exit channel (5, 6, 16). Structure-based models predict steric clash between the nascent transcript and σ region 4 when the nascent RNA first emerges from the RNA exit channel (6) at a length of ≈15–16 nt (19, 20). It has therefore been proposed that σ region 4 must be dislodged from the β-flap to permit extension of the nascent RNA beyond this point (6). Predictions arising from this proposal are that weakening the σ⁷⁰ region 4/β-flap interaction should facilitate extension of the nascent RNA beyond ≈15–16 nt, and conversely, that strengthening this interaction should inhibit extension of the nascent RNA beyond ≈15–16 nt. To test these predictions, we took advantage of an early elongation pause that occurs during transcription from the bacteriophage λ late promoter P_R′ (34). At this promoter, RNAP holoenzyme pauses as a result of an interaction between region 2 of σ⁷⁰ and a DNA sequence element in the initial transcribed region that resembles a promoter –10 element (35). The pause, which is strictly dependent on the presence of σ⁷⁰ in elongating RNAP, is manifested in early elongation complexes containing a 16- or 17-nt transcript (34). In particular, a certain fraction of paused early elongation complexes at λP_R′ contains a 16-nt transcript, and a certain fraction contains a 17-nt transcript.

We therefore sought to take advantage of the amino acid substitutions that we had isolated to ask whether altering the strength of the σ⁷⁰ region 4/β-flap interaction would alter the proportion of paused early elongation complexes containing either a 16- or 17-nt nascent transcript. To address this question, we performed single-round in vitro transcription assays using Eσ⁷⁰ WT, Eσ⁷⁰ L607P, or Eσ⁷⁰ T544I/D581G (Fig. 5A). When reactions were performed with Eσ⁷⁰ WT, among the fraction of RNAP molecules that paused, the ratio of paused elongation complexes containing a 17-nt transcript to paused elongation complexes containing a 16-nt transcript (+17/+16 ratio) was 1.9 (± 0.2) (Fig. 5B). In contrast, when reactions were performed with Eσ⁷⁰ L607P (a substitution that weakens the σ⁷⁰ region 4/β-flap interaction), the +17/+16 ratio was 5.2 (± 0.7), and when reactions were performed with Eσ⁷⁰ T544I/D581G (substitutions that strengthen the σ⁷⁰ region 4/β-flap interaction), the +17/+16 ratio was 1.0 (± 0.1). We conclude from these results that weakening the σ⁷⁰ region 4/β-flap interaction increases the probability that RNAP will add the 17th nt to the nascent RNA in the context of an early elongation complex at λP_R′. Conversely, strengthening the σ⁷⁰ region 4/β-flap interaction decreases the probability that RNAP will add the 17th nt to the nascent RNA.

Discussion

We report the isolation of amino acid substitutions in σ⁷⁰ region 4 that specifically alter the strength of its interaction with the flap domain of the RNAP β-subunit. Characterization of RNAP holoenzymes containing the corresponding σ⁷⁰ mutants reveals that altering the strength of the σ⁷⁰ region 4/β-flap interaction affects not only transcription initiation but also transcription elongation.

σ⁷⁰ Region 4/β-Flap Interaction in Transcription Initiation: Recognition of Extended –10 Promoters. We found that weakening the interaction between σ⁷⁰ region 4 and the β-flap resulted in defects in the utilization not only of –10/–35 promoters but also of a consensus extended –10 promoter (galP1/cons) (Fig. 4; see also Note 1 in Supporting Text, which is published as supporting information on the PNAS web site). Although galP1/cons lacks a recognizable –35 element, it has been shown previously that, in the RNAP-promoter open complex at galP1/cons, σ⁷⁰ region 4 interacts with DNA in the –35 region (9, 36, 37). The σ⁷⁰/β-flap interaction may help position σ⁷⁰ region 4 for such nonspecific interaction with DNA in the –35 region. Alternatively, or in addition, the σ⁷⁰/β-flap interaction may help position σ⁷⁰ region 4 for possible interaction with the α-C-terminal domain (38, 39). One or both of these interactions could stabilize the binding of RNAP to the promoter and/or facilitate promoter melting (36).

A previous study (9) reported that an RNAP derivative lacking 30 residues of the β-flap is defective in transcription initiation from a –10/–35 promoter (T7A2), but not from galP1/cons. Possibly, the difference between this previous result and our finding here reflects a functional difference between the holoenzyme in which the β-flap has been deleted and holoenzymes in which the σ⁷⁰ region 4/β-flap interaction has been destabilized by single or double amino acid substitutions. Another previous study (33) identified single amino acid substitutions in the β-flap-tip helix that preferentially inhibited promoter engagement at a –10/–35 promoter; however, smaller effects on promoter engagement at galP1/cons were not excluded.

We also found that weakening the σ⁷⁰ region 4/β-flap interaction causes a selective loss of a prominent abortive RNA product that is synthesized at galP1/cons. It has been proposed that the clash between the nascent RNA and σ⁷⁰ region 3.2 positioned in the RNA exit channel contributes to abortive initiation (5, 24); the extent of abortive initiation can also be influenced by promoter strength (40). Destabilization of the σ⁷⁰ region 4/β-flap interaction could result in a reduction in abortive RNA products because σ⁷⁰ region 3.2 may be more easily displaced from the RNA exit channel and/or because of the loss of sequence-nonspecific interactions between σ⁷⁰ region 4 and DNA in the –35 region.

σ⁷⁰ Region 4/β-Flap Interaction in Transcription Elongation: Extension of 16-nt Nascent RNA. We observed an inverse correlation between the relative strength of the σ⁷⁰ region 4/β-flap interaction and the ability of RNAP holoenzyme to extend the nascent RNA past position 16 when σ⁷⁰ region 2 is bound to the pause-inducing –10-like element at λP_R′. In light of structural considerations, we propose that the effect of the σ⁷⁰ region 4/β-flap interaction at λP_R′ is the manifestation of steric clash between the nascent RNA and region 4 that occurs when the nascent RNA first emerges from the RNA exit channel, at a length of ≈15–16 nt (19, 20). In particular, we propose that the nascent RNA must displace region 4 of σ⁷⁰ from the β-flap, or at least alter the interaction between region 4 and the β-flap, for elongation to proceed past 16 nt (Fig. 6). Thus, when the interaction between σ⁷⁰ region 4 and the β-flap is weakened, extension past 16 nt is favored; and, when the interaction between σ⁷⁰ region 4 and the β-flap is strengthened, extension past 16 nt is disfavored. We emphasize that at λP_R′, the inferred steric clash between the nascent RNA and region 4 of σ⁷⁰ manifests itself only as a redistribution of paused elongation complexes (see Supporting Text, Notes 2 and 3), but is not required for their formation (that depends strictly on the protein/DNA interaction of σ⁷⁰ region 2 and the pause-inducing sequence element). In the case of λP_R′, this clash evidently occurs at a nascent-RNA length of 16 nt, but the exact point of the clash might vary, depending on the sequence of the initial transcribed region; in addition, the clash might in some cases occur before promoter escape (40).

Fig. 6.

Extension of nascent RNA past 16 nt at λP_R′ requires displacement of σ⁷⁰ region 4 from the β-flap. Cartoons depict sequence of events during initiation and early elongation at λP_R′. (Upper) The initiation complex. σ⁷⁰ region 3.2 (σ_3.2) is located within the RNA exit channel and σ⁷⁰ region 4 (σ₄) is bound to the β-flap. (Lower) The paused elongation complex with σ⁷⁰ region 2(σ₂) bound to the pause-inducing –10-like element. The nascent RNA, which is not shown, has displaced σ⁷⁰ region 3.2 from the RNA exit channel. Enlargements depict proposed clash of the nascent RNA with σ₄ bound to the β-flap that requires displacement of σ₄ for addition of the 17th nt.

Staged Disruption of σ⁷⁰/Core Interactions in Transition from Initiation to Elongation. Our finding that σ⁷⁰ region 4 presents a barrier to the nascent RNA as it emerges from the RNA exit channel supports the view that the transition from initiation to elongation involves staged displacement of segments of σ⁷⁰ from RNAP core (4–6, 16, 18, 24). According to this view, the nascent RNA displaces distinct segments of σ⁷⁰ in distinct steps. In a first step, entry of the nascent RNA into the RNA exit channel, at a length of 10–11 nt, displaces σ⁷⁰ region 3.2 from the RNA exit channel, and possibly, as a consequence, weakens the σ⁷⁰ region 4/β-flap interaction. In a second step, emergence of the nascent RNA from the RNA exit channel, at a length of 15–16 nt, displaces σ⁷⁰ region 4 from the β-flap, which may serve to prevent σ⁷⁰ region 4 from making nonspecific interactions with the DNA that would impede the elongation process. Our findings provide support for the existence of the second step and suggest that displacement of σ⁷⁰ region 4 from the β-flap is an obligatory step during formation of a mature elongation complex.