The interaction between bacterial transcription factors and RNA polymerase during the transition from initiation to elongation

Abstract

There are three stages of transcription: initiation, elongation and termination, and traditionally there has been a clear distinction between the stages. The specificity factor sigma is completely released from bacterial RNA polymerase after initiation, and then recycled for another round of transcription. Elongation factors then associate with the polymerase followed by termination factors (where necessary). These factors dissociate prior to initiation of a new round of transcription. However, there is growing evidence suggesting that sigma factors can be retained in the elongation complex. The structure of bacterial RNAP in complex with an essential elongation factor NusA has recently been published, which suggested rather than competing for the major σ binding site, NusA binds to a discrete region on RNAP. A model was proposed to help explain the way in which both factors could be associated with RNAP during the transition from transcription initiation to elongation.

The first step in gene expression, transcription, is one of the most highly regulated processes in the cell. In all organisms, transcription is carried out by a DNA-dependent RNA polymerase (RNAP). In eukaryotic cells, there are three distinct RNAPs: RNAP I transcribes rRNA precursors, RNAP II is responsible for the synthesis of mRNA and certain small nuclear RNAs and the synthesis of 5s rRNA and tRNA is performed by RNAP III.1 In bacteria, there is only one RNAP responsible for the transcription of all classes of RNA. The evolutionarily conserved ∼400 kDa bacterial RNAP core enzyme (subunit composition α₂ββ'ω) is capable of catalyzing RNA synthesis on its own. However, to initiate transcription RNAP core must be associated with a σ factor forming a holoenzyme, which is competent for specific binding to the promoter regions on DNA.2 A typical promoter generally comprises two hexanucleotide consensus motifs: the −10 and −35 elements (nucleotide sequence: TTGACA and TATAAT, respectively) which are separated by a ∼17 base pair (bp) spacer.3 Most bacteria have one primary housekeeping σ factor (σ⁷⁰ in Escherichia coli and σ^A in Bacillus subtilis) and one or a number of alternative σ factors for transcribing specific classes of genes.4 σ⁷⁰-family members are both sequentially and structurally conserved among all bacteria and have four regions of homology designated 1–4, which can be further subdivided into evolutionarily conserved regions.5 A DNA-binding domain of σ region 2, σ_2.4, is responsible for sequence-specific interactions with the non-template strand of the −10 promoter element and σ factor region 4.2 (σ_4.2) binds to the −35 promoter element.6 The major RNAP-σ interface is formed between the region 2.2 of σ (σ_2.2) and the solvent exposed clamp-helix (CH) region of the RNAP β' subunit (Fig. 1A).7,8 Region 4 of the σ factor (σ₄) also forms weak contacts with the flap domain (β flap) of the β subunit of RNAP, immediately adjacent to the tip of the RNA exit channel (Fig. 1A).9

There are three main sequential steps in the transcription cycle: promoter binding/initiation, RNA chain elongation and termination. Each step or intermediates within steps in the cycle represent a checkpoint for the regulation of gene expression by transcription factors.10 To start transcription, RNAP scans the DNA (Fig. 1B) until a specific promoter region is recognized by the σ factor with σ_2.4 binding to the −10 promoter element and σ_4.2 binding to the −35 element. Subsequently ∼14 bp of DNA are melted around the AT-rich −10 box and then RNA synthesis is initiated at the +1 start site.11 In most cases RNAP remains at the promoter region and initiates several rounds of abortive initiation, generating short transcripts 2–15 nucleotides (nt) in length.12 Once about 12 nt of RNA have been synthesized, it is sufficiently long to fill the upstream RNA exit channel and destabilize the binding of σ region 4 to the β flap (Fig. 1C). Disruption of the σ₄-β flap interaction is the first step of σ dissociation from RNAP9 and in turn triggers the release of the −35 promoter element. This allows RNAP to escape from the promoter and enter the elongation phase. It is widely accepted that after initiation factor σ completely dissociates from RNAP core, elongation factors become associated to regulate elongation and termination, followed by rebinding of σ to allow a new round of transcription initiation (the σ cycle).13

While the σ cycle is likely to be true most of the time, there is also some evidence that challenges the traditional view. σ factor has been shown to be present in a proportion of elongation complexes (ECs)14,15 and possibly plays functional roles in regulating transcription elongation (reviewed in ref. 13). The best-characterized system involves the formation of the bacteriophage λ Q protein antitermination complex required for the regulation of late gene transcription from the λ late promoter P_R' (Fig. 2). There are two cis-acting sequences within the P_R' regions required for the Q protein to engage the transcription machinery during the transition between initiation and elongation: a Q binding element (QBE) and a pause-inducing element (Fig. 2A). As the nascent RNA transcribed from the P_R' promoter emerges from the RNA exit channel at a length of ∼16 nucleotides, σ₄ region is released from the β flap region and σ₂ binds to the pause-inducing element (which resembles a promoter −10 element) and triggers pausing of the early EC (Fig. 2B). The Q protein interacts with the QBE located between the promoter −10 and −35 elements, and engages this paused EC around the β flap region of RNAP. The Q protein also interacts with σ₄ and stabilizes its binding to the −35-like pause-inducing element (Fig. 2C). The σ factor is thus retained after RNAP has escaped the promoter and adopts a conformation in which both region 2 and 4 simultaneously bind the pause-inducing element (Fig. 2C). This permits the Q protein to become a stable component of the EC, resulting in the formation of an antitermination complex, allowing RNAP to read through downstream transcription terminators and transcribe the phage λ late genes.16,17

The λ system exemplifies how σ factors are involved in the recruitment of elongation factors to RNAP. Another example we are focusing in this review is the interaction of the σ factor and elongation factor NusA with RNAP during the transition between initiation and elongation. NusA is highly conserved and essential in many organisms. It was first identified as the host factor involved in phage λ N-protein mediated antitermination18 and is also required in the antitermination complexes involved in host cell rRNA synthesis.19 Conversely, NusA enhances RNAP pausing during transcription, particularly at pause sites, to ensure efficient intrinsic transcription termination.20 NusA has also been shown to be involved in the “immune system” to suppress the toxic activity of foreign genes in E. coli host cells.21 Structurally, NusA is a highly elongated molecule with the N-terminal domain (NTD) required for RNAP interaction joined by a flexible linker to the C-terminal S1/KH RNA binding domains.22,23

The traditional view of the transcription cycle involves the release of σ as RNAP clears the promoter, coincident with the binding of NusA and entry into the elongation phase of transcription. Results from several studies also showed that the binding of NusA and σ to RNAP was mutually exclusive, i.e., NusA cannot interact with RNAP when σ is bound and vice versa.24 An α-helical region of the NusA NTD has also been observed to share structural homology with the major RNAP binding domain of σ factor (σ₂), suggesting that the mutual exclusivity of binding resulted from direct competition of both factors for the same binding site on RNAP.25 However, in vitro and in vivo results from other studies involving affinity isolation, single molecule and ChIP-chip analysis are inconsistent with the model due to the fact that both σ and NusA can be present simultaneously in transcription complexes under certain circumstances.14,15,26,27 Recently we proposed a model for RNAP in complex with NusA based on results from biochemical mapping and structural approaches.28 Rather than competing for the major σ factor binding site (CH region of β'), NusA has been shown to bind principally to the β flap region via its NTD with the C-terminal domain (CTD) wrapped back across the β flap where it interacts with the emerging transcript.28 The proposed NusA-RNAP complex model, which is consistent with other previous biochemical observations, has helped us better understand the various roles of NusA in the transcription cycle (also reviewed in ref. 29) and also enables the traditional σ cycle to be carefully re-examined.

In the initiation complex the σ_2.4 and σ_4.2 bind to the −10 and −35 promoter element respectively and in this conformation σ_2.2 binds to the CH domain of the β' subunit and σ₄ to the β flap (Fig. 1C). The σ₂-CH interaction has been shown to contribute to the majority of σ factor's affinity to RNAP, which could in turn stabilize the binding of σ₄ to the β flap (Johnston EB and Lewis PJ; unpublished). Thus, NusA will not be able to compete with σ₄ for its binding site on the β flap.28 This is consistent with the fact that σ⁷⁰ showed a much higher affinity for RNAP during initiation than NusA in a competition assay.30 Once the promoter is cleared and σ₄ is released from the β flap, NusA is able to become associated with RNAP at the β flap region via its NTD (Fig. 1A and D). Since the interactions between σ_2.2 and the β' CH region of RNAP are not directly affected by σ₄ displacement or NusA binding, a small fraction of σ could be retained. As a result, a transition complex from the initiation to elongation phase of transcription is formed containing σ bound to the CH region and NusA bound around the β flap, which has been previously reported (Fig. 1D).14 This is also consistent with recently published genome wide chromatin immunoprecipitation-microarray hybridization (ChIP-chip) data that shows that as σ slowly dissociates from the initiation complex after RNAP clears the promoter, NusA becomes associated, before the σ signal fully disappears.27 Binding of NusA prevents σ₄ from rebinding to the σ flap region, and also is likely to induce structural changes in RNAP resulting in the formation of a stable transcription EC, which in turn further reduces the affinity of σ interaction to RNAP, as observed previously.30 In most cases, σ is gradually released from RNAP during elongation (Fig. 1E). However, since the σ_2.2 and β' CH interaction is not affected by NusA binding, under some circumstances σ could remain associated with (or rebind to) RNAP during elongation or through multiple rounds of transcription, as previously proposed.13,14 This is consistent with the ChIP-chip data that shows that a small proportion of the transcription complexes appear to retain σ throughout the cycle.27

In summary, whilst the classic σ cycle, where σ completely dissociates from RNAP after initiation and rebinds for a new round of transcription, is likely to occur most of the time, there have been studies suggesting that this may not always be the case. σ factor has been shown to remain associated with the ECs and play regulatory roles under some circumstances. The recently published model of RNAP in complex with NusA has allowed us to refine the classic σ cycle and make an informed hypothesis on how the factors are interacting with RNAP during the transition from initiation to elongation.

Abbreviations

bp	=	base pair
CH	=	clamp helix
CTD	=	C-terminal domain
ChIPchip	=	chromatin immunoprecipitation-microarray hybridization
EC	=	elongation complex
NTD	=	N-terminal domain
QBE	=	Q binding element
RNAP	=	RNA polymerase

Figures and Tables

Figure 1 (A) Structural model of RNAP with the relative positions of the α, β, β' and ω subunits marked. The boxed areas indicate the clamp-helix (CH) region and β-flaps which are shown expanded below. σ region 2 binds the CH region, whilst σ region 4 and NusA NTD bind the β-flap. (B–E) Cartoon representation of the transcription machinery at different stages. (B) RNAP scans along the DNA until it encounters a promoter. (C) After a few rounds of abortive initiation, σ factor region 4 dissociates from the β flap, while region 2 remains bound to the CH region. (D) NusA binds to RNAP via its NTD while the rest of σ remains associated. (E) In most cases, σ is released from RNAP, resulting in a stable EC. See main text for detailed discussion.

Figure 2 Formation of the phage Q protein antitermination complex at the P_R' promoter. (A) The initiation complex with ₂ bound to the −10 promoter element and 4 to the −35 promoter element. ₄ is also shown interacting with the β flap (triangle) of RNAP. (B) The paused early EC at P_R' with ₂ bound to the pause-inducing −10-like element. 4 (dotted outline) is released from the β flap. (C) Q (shown as a dimer) binds to the QBE and become associated with the paused EC at the β flap region. Q also interacts with ₄ and stabilizes its binding to the −35-like pausing element. Adapted from Nickels et al (ref. 16). See main text for detailed discussion.

Acknowledgements

Work in the lab of P.L. is supported by the ARC and NHMRC. X.Y. was supported by an Australian Postgraduate Award from the Australian Government.