• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptNIH Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Mol Cell. Author manuscript; available in PMC Jan 16, 2010.
Published in final edited form as:
PMCID: PMC2747249

Regulator trafficking on bacterial transcription units in vivo


The in vivo trafficking patterns on DNA by the bacterial regulators of transcript elongation σ70, ρ, NusA, and NusG and the explanation for high promoter-proximal levels or peaks of RNA polymerase (RNAP) are unknown. Genome-wide ChIP-chip on E. coli revealed distinct association patterns of regulators as RNAP transcribes away from promoters (ρ first, then NusA, and then NusG). However, the interactions of elongating complexes with these regulators, including a weak interaction with σ70, did not differ significantly among most transcription units. A modest variation of NusG signal among genes reflected increased NusG interaction as transcription progresses, rather than functional specialization of elongating complexes. Promoter-proximal RNAP peaks were offset from σ70 peaks in the direction of transcription and co-occurred with NusA and ρ peaks, suggesting that the RNAP peaks reflected elongating, rather than initiating, complexes. However, inhibition of ρ did not increase RNAP levels within genes downstream of the RNAP peaks, suggesting the peaks are caused by a mechanism other than simple ρ-dependent attenuation.


Transcription of genes by RNAP is controlled by a multiplicity of regulators that modulate template DNA conformation, control initiation, or govern RNAP’s progress through transcription units (TUs) in response to internal and environmental signals. In bacteria and eukaryotes, transcription regulators can be divided into those acting during transcript initiation, elongation, or termination. Precisely where initiation regulators release and elongation regulators associate with RNAP is unknown. Further, the distinction between these classes of regulators is not absolute; some may act during multiple stages of transcription, possibly with different effects. Finally, although some elongation regulators are known to target subsets of TUs, it is unclear whether general elongation regulators like NusA, NusG, and ρ interact with most elongating complexes (ECs) equivalently or instead preferentially interact with certain TUs or sites within TUs.

In bacteria, σ initiation factors bind tightly to core RNAP (consisting of β’, β, α2, and ω subunits) and determine the sequence specificity of RNAP-promoter interactions (Fig. 1A). σs are thought to be released shortly after RNA synthesis begins. However, whether σ release occurs obligately or stochastically, whether σ may be completely retained on a subset of TUs, and whether σ may transiently rebind to the EC during elongation with possible regulatory consequence all remain in debate (Bar-Nahum and Nudler, 2001; Kapanidis et al., 2005; Mooney et al., 2005; Mooney and Landick, 2003; Mukhopadhyay et al., 2001; Raffaelle et al., 2005; Reppas et al., 2006; Wade and Struhl, 2004, 2008).

Fig. 1
Bacterial regulators of transcript elongation

During or after promoter escape, the EC can associate with one or more elongation regulator (Fig. 1A). In bacteria, NusA and NusG alter EC properties differently via direct and independent interactions with RNAP, and are the best characterized regulators of elongation (Burns et al., 1998; Greenblatt et al., 1981; Li et al., 1992; Linn and Greenblatt, 1992; Sullivan and Gottesman, 1992). NusA preferentially enhances transcriptional pausing associated with nascent RNA hairpins (Artsimovitch and Landick, 2000; Farnham et al., 1982; Greenblatt et al., 1981; Yakhnin and Babitzke, 2002), enhances intrinsic termination at some sites more than others (Kassavetis and Chamberlin, 1981; Linn and Greenblatt, 1992; Yakhnin and Babitzke, 2002), modulates ρ-dependent termination (Burns et al., 1998), and is an essential component of antitermination complexes that form on ribosomal RNA (rrn) and phage λ operons (Mason et al., 1992; Shankar et al., 2007; Torres et al., 2001). NusG increases the rate of RNA chain extension, at least partly by decreasing pausing associated with backtracking (Artsimovitch and Landick, 2000), enhances ρ-dependent termination via interactions with RNAP and ρ (Li et al., 1992, 1993; Sullivan and Gottesman, 1992), and also is a component of both rrn and λ antitermination complexes (Mason et al., 1992; Torres et al., 2001). Despite these multiple roles of NusA and NusG, it is unclear whether they associate equivalently with ECs on all TUs, differentially with subsets of TUs, or differentially at locations within TUs.

The homohexameric ρ protein terminates transcription after binding to unstructured, C-rich nascent RNA. RNA-stimulation of its ATP-dependent translocase activity allows ρ to travel 5′ to 3′ along the RNA and dissociate ECs unless blocked by intervening ribosomes (Richardson, 2002). It is uncertain where within TUs ρ interacts with ECs and whether ρ preferentially affects a subset of TUs. The report of Reppas et al. (2006) that a significant fraction of TUs in E. coli exhibit promoter-proximal peaks of RNAP heightens interest in knowing whether promoter-proximal, ρ-dependent termination could contribute to the apparent decrease in RNAP density downstream from promoters.

To investigate trafficking of these regulators on bacterial TUs and the reported promoter-proximal block to transcription (Reppas et al., 2006), we used “chromatin immunoprecipitation” (Kuo and Allis, 1999; Solomon et al., 1988) followed by microarray hybridization (ChIP-chip; Wade et al., 2007). Our study provides comparative analysis with improved resolution of some proteins examined previously (RNAP, σ70, and NusA; Grainger et al., 2005; Herring et al., 2005; Raffaelle et al., 2005; Reppas et al., 2006; Wade and Struhl, 2004) as well as the first genome-wide views of NusA, NusG and ρ, leading to important new insights into trafficking of bacterial transcription regulators.


Analysis of RNAP ChIP-chip signals on E. coli TUs

We applied ChIP-chip to E. coli K-12 at mid-log phase of growth at 37 °C in defined minimal glucose medium (Experimental Procedures), conditions in which many biosynthetic genes must be expressed and that were used previously for expression analysis (Allen et al., 2003). Using specific antibodies targeting core RNAP, σ70, NusA, ρ, or a hemagglutinin (HA) epitope present in three copies at the N-terminus of the chromosomal nusG gene, we obtained associated DNA that was then fluorescently labeled and hybridized to a tiled oligonucleotide microarray (~25 bp spacing; Experimental Procedures). Initial analysis of the immunoprecipitated DNAs relative to input DNA revealed excellent correspondence among the sites of enrichment by anti-σ70 and anti-RNAP (anti-β’) antibodies (Fig. 1B). Closer examination (e. g., of the expanded region around 0.94 mb shown in Fig. 1B) revealed that σ70 was predominantly associated with DNA near promoters, whereas RNAP could be detected in association with both promoter and transcribed-region DNA. The strongest signals were in genes encoding tRNA, rRNA, and ribosomal proteins (e. g., serW and rpsA), as expected and reported previously (Grainger et al., 2005; Raffaelle et al., 2005; Reppas et al., 2006; Wade and Struhl, 2004). NusA, NusG, and ρ were associated with ECs in most locations where RNAP was present.

RNAP is known to associate non-specifically with chromosomal DNA (deHaseth et al., 1978; Grigorova et al., 2006; von Hippel et al., 1974). To estimate the corresponding non-specific (background) ChIP-chip signal for RNAP, we examined regions of the bacterial chromosome thought to be devoid of transcription, such as the cryptic bglB gene (Defez and De Felice, 1981). We identified 170 regions greater than 1 kb whose average RNAP ChIP signal was indistinguishable from that on bglB (bkgd, Fig. 1B; gray box near 0.94 mB in expanded region; Table S2). The signals for these regions were normally distributed with a mean below the signal for ~84% of the complete genome-wide probe set (compare black to blue histograms, Fig. 1C; Supplemental Experimental Procedures). This suggests that most of the E. coli genome is transcribed at levels above the non-specific background, consistent with previous estimates (Selinger et al., 2000).

To characterize RNAP and regulator occupancy further, we identified “high-quality” TUs that were significantly above this background and for which signals from adjacent TUs did not obscure the pattern of RNAP and regulator association and dissociation (e.g., serS in the expanded region as opposed to clpA and cydCD, which were obscured by strong signals from the adjacent serW tRNA gene). We identified 109 such TUs, which were spread across the E. coli genome and represented a range of expression levels and TU lengths (Fig. 1D and Table S1).

Regulator trafficking on representative E. coli TUs

To gauge the basic patterns of regulator trafficking on these 109 TUs, we wished to scale the data in proportion to occupancy of regulators on DNA. Although true occupancy is impossible to measure without knowing the relative efficiencies of crosslinking for each protein at each TU location as well as the signals corresponding to zero and full occupancy, we nevertheless defined an apparent occupancy (Occapp) by linearly scaling signals for each protein between zero, which was set equal to the background defined by bglB-similar regions (Fig. 1C; Table S2), and one, which was arbitrarily defined as the average of the ten 3-probe clusters with highest average value (Fig. 1C; Supplemental Experimental Procedures). Therefore, Occapp is a function of true occupancy and relative “crosslinkability.”

An examination of eight representative TUs (seven from among the 109 high-quality TUs plus rrnE) revealed significant variation both in the uniformity of RNAP and regulator Occapp across TUs and in the ratios of RNAP Occapp to σ70 and other regulators at locations within TUs (Fig. 2). In some cases, the peak of σ70 Occapp surrounding the transcription start site (TSS) was much greater than RNAP Occapp, with the latter exhibiting a relatively uniform distribution across the TU (serS, rspF, and acnB; Figs. 2A, D, and F). In other cases, the σ70 peak was more similar to the corresponding RNAP Occapp (atpIBEFHAGDC, gltBDF, and carAB; Figs. 2C, E, and H); in these cases RNAP typically exhibited a pronounced promoter-proximal peak similar to that previously reported (Reppas et al., 2006; Wade and Struhl, 2008). These representative examples suggest there is no one-to-one correspondence between σ70 Occapp and RNAP Occapp at promoters; this observation was reflected in the modest (0.77) correlation between peak Occapp values for σ70 and RNAP (Fig. S1).

Fig. 2
Apparent occupancy profiles of RNAP and regulators on representative TUs

σ70, NusA, NusG, and ρ associate with ECs in different patterns

The regulators σ70, NusA, NusG, and ρ all appeared to be present on each TU, but with notable differences in their Occapp distributions. NusA cosely mirrored RNAP on each TU, appearing to associate with RNAP as the signal from σ70 disappears. This is consistent with the long-standing view that NusA displaces σ70 during transcript elongation (Greenblatt and Li, 1981). In contrast, NusG appeared to associate with elongating RNAP farther from promoters and did not appear to be present at locations where RNAP forms promoter-proximal peaks. Rather, NusG Occapp rose gradually to levels that exceed other regulators on most TUs. The ratio of NusG/RNAP Occapp appeared to be much greater in the distal portions of some TUs (e.g., atpIBEFHAGDC and cyoABCDE; Figs. 2C and G) than others (e.g., rrnE and rpsF-priB-rpsR-rplI; Figs. 2B and D). The different pattern of NusG on rrnE may reflect its participation (with NusA, NusB, NusE, and a subset of ribosomal proteins) in the rrn antitermination complex (Torres et al., 2001). ρ exhibited a striking pattern of significant promoter-proximal peaks near σ70 peaks and RNAP promoter-proximal peaks, but a lower Occapp over most of the TU. Finally, although σ70 was principally present at promoters, as reported previously (Reppas et al., 2006; Wade and Struhl, 2004), σ70 Occapp remained above zero across most TUs (e.g., serS, rrnE, and cyoABCDE; Figs. 2A, B, and G).

To examine the correlation between RNAP and regulator presence on TUs more carefully, we calculated the average ChIP-chip signals for each in a 200 bp window in the middle of the 109 high-quality TUs (Fig. 3A) and compared the regulator and RNAP ChIP-chip signals directly (Figs. 3B-F). Strikingly, σ70, NusA, NusG, and ρ mid-TU signals all exhibit an obvious correlation with RNAP mid-TU signals. However, the correlation was much greater for NusA than for σ70, NusG, or ρ (Fig. 3F). For σ70 and ρ, the weaker correlation is consistent with lower signal-to-noise ratio resulting from the reduced mean signals in the middle of the TUs. However, this is not the case for NusG, where the mean mid-gene signal was as large as the RNAP signal despite the much-reduced correlation (Fig. 3F). These results suggest that elongating RNAPs do not exhibit TU-specific variations in affinity for σ70, NusA, NusG, or ρ. Although the relative affinity of each regulator for ECs differs (i.e., σ70 and ρ exhibit lower signals than NusA and NusG), there is no indication that they target one subset of TUs relative to others. Thus, they can rightly be classified as general elongation regulators as opposed to specialized regulators like RfaH that are recruited to a specific subset of TUs (Artsimovitch and Landick, 2002).

Fig. 3
Mid-TU regulator signals correlated with RNAP signals

To resolve the pattern of σ70, NusA, NusG, and ρ interactions with RNAP more accurately, we took advantage of the similarity of these interactions among TUs to compute aggregate Occapp profiles (Fig. 4). For this purpose, we selected a set of highly transcribed TUs among the 109 high-quality TUs (to improve signal-to-noise ratios) and avoided TUs known to contain transcription attenuators (e.g., trp or leu) or multiple promoters that might complicate the distribution of RNAP. This yielded a set of 42 TUs that included 13 lacking an obvious promoter-proximal RNAP peak and 29 containing a readily discerned promoter-proximal RNAP peak (traces B and C in Fig. 4A). We computed the aggregate Occapp for these TUs by aligning them relative to the genome coordinate of their σ70 peak and then averaging normalized Occapp values for each protein (normalized relative to the highest Occapp for that protein in a given TU). The RNAP peak aggregate Occapp for the 42 TUs was offset in the direction of transcription from the σ70 peak by ~150 bp (d in Fig 4A). The size of this offset was widely distributed among different TUs and was uncorrelated with RNAP mid-TU signal (Fig. S2). However, the 29 TUs exhibiting pronounced peaks were, on average, longer (3.43 kb average length), whereas the TUs on which Occapp declined much more slowly were, on average, shorter TUs (1.36 kb average length; Mann-Whitney p<0.001).

Fig. 4
Aggregate apparent occupancy for highly expressed TUs

The aggregate Occapp profiles highlighted differences in regulator trafficking on E. coli TUs. σ70 appeared to dissociate from RNAP as RNAP loses contact with the promoter (as reported previously by Raffaelle et al., 2005; Reppas et al., 2006; Wade and Struhl, 2004). Although the σ70 peak was nearly symmetric around its center as noted by Reppas et al. (2006), it was skewed ~20 bp downstream at its vertical midpoint in our data (Fig. S3). This σ70 skew was caused by translocation of RNAP relative to the TSS, as evidenced by loss of the skew and a slight upstream shift of the σ70 peak upon treatment of cells with rifampicin (Fig. S3). Conversely, NusA appeared to associate fully with elongating RNAP sometime after the σ70 signal disappeared (Figs. 4B and C). Both the NusA and ρ aggregate profiles exhibited promoter-proximal peaks, as observed for the individual profiles (compare Figs. Figs.22 and and4C).4C). However, the ρ peak was displaced ~50 bp upstream (relative to the RNAP peak), whereas the NusA peak was displaced downstream. Finally, NusG associated with elongating RNAP much more slowly than either NusA or ρ, reaching a plateau of Occapp ~800 bp downstream of the σ70 peak. The same aggregate and individual-TU patterns of NusG association were observed using anti-NusG polyclonal antibody (Fig. S4), ruling out perturbation caused by the HA3 tag.

Taken together, our analysis of regulator trafficking on E. coli TUs (Figs. (Figs.22--4)4) leads to the following key conclusions. First, σ70 crosslinks almost exclusively to promoter DNA, although a downstream skew of the σ70 peak and weak σ70 signal in the middle of TUs are consistent with stochastic release of σ70 from elongating RNAP followed by weak σ70 association with ECs (Mooney et al., 2005). The extent of σ70-EC association is difficult to assess from ChIP-chip data (see Discussion); we cannot exclude the possibility that non-specific antibody-EC interaction contributes to the mid-TU σ70.

Second, NusG associates with ECs more slowly than NusA on most TUs (Figs. (Figs.22 and and4),4), except on antiterminated rrn TUs where its faster association likely reflects incorporation into an antiterminated EC (Torres et al., 2001). Conversely, the slower association of NusG on other TUs may suggest its binding is stimulated by a feature of the EC that increases the farther RNAP transcribes.

Third, ρ is evident at most TU locations, with a peak interaction at locations in between the strongest σ70 and RNAP signals (Figs 4B and C). This suggests that ρ may associate with transcripts shortly after the initiation of transcription. ρ is detectable throughout TUs, and the extent of this interaction is well-correlated with the amount of RNAP located on the TU (Fig. 3E). This is consistent with the generally accepted role of ρ in premature termination whenever translation is compromised.

NusG apparent occupancy depends on TU length, not gene function

To investigate the greater variability of NusG/RNAP ratios and NusG’s apparently slower association with ECs, we computed the NusG/RNAP, NusA/RNAP, and ρ/RNAP ratios for each gene and examined these ratios as a function of the average RNAP signal per gene (Figs. 5A-C). NusA and ρ both exhibited relatively uniform distributions; genes with low RNAP signals exhibited higher ratios (as expected mathematically; Figs. 5A-B). In this analysis, NusA/RNAP ratios on rRNA genes were slightly above the trend line, but were still consistent with at least 1:1 NusA:RNAP on most ECs. tRNA genes exhibited disproportionately high ratios of both NusA and ρ, suggesting transcription of tRNA genes may differ from protein-coding genes. Small RNA (sRNA) genes, in contrast, exhibited normal ratios of NusA and ρ to RNAP.

Fig. 5
Gene-averaged regulator/RNAP ratios

The NusG/RNAP ratio distribution differed strikingly from the NusA or ρ ratios. Although rRNA genes exhibited high NusG/RNAP ratios, a subset of genes with lower average RNAP signal exhibited even higher NusG/RNAP ratios (inset, Fig. 5C). Interestingly, several of these were genes involved in energy production (genes from the nuo and cyo operons), murein/peptidoglycan biosynthesis and recycling (oppD&F, murB&E), or amino-acid biosynthesis (trpA&B, metI, cysM). This raised the possibility of a functional connection to elevated NusG levels on certain TUs (e.g., to localize transcription of certain genes). As an alternative, we considered whether the length of TUs might explain the abnormal NusG/RNAP ratios (e.g., if long TUs acquire higher NusG occupancy). To test this, we compared the NusG/RNAP ratio to the distance of genes from their TSS (for cases where the TSS is known) and found a strong correlation of TSS-gene distance to NusG/RNAP ratio (Spearman r = 0.57; Fig. 5D). Genes that deviated significantly from this strong correlation by exhibiting low NusG/RNAP ratios included rfa and rfb genes (inset, Figure 5D). This is readily explained because rfa and rfb genes are regulated by RfaH, a specialized paralog of NusG that competes with NusG for interaction with ECs (Belogurov et al., 2007).

We conclude that the gradual increase in NusG association as transcription progresses, rather than a connection to gene function, explains elevated NusG/RNAP ratios on some genes. The high NusG/RNAP ratios on energy-related and amino-acid-biosynthetic operons simply reflect the greater-than-average length of these TUs. To confirm this interpretation, we plotted the average NusG/RNAP ratios for different gene functional classes by the average TSS-gene distance for the functional class (Fig. 5E). Classes with NusG/RNAP signal ratios below the genome average (red circle, Fig. 5E) contained, on average, shorter genes, whereas classes exhibiting significantly higher NusG/RNAP signal ratios contained longer genes. Thus, the primary determinant of NusG levels is TSS-gene distance, rather than gene function.

Promoter-proximal RNAP peaks correlate with promoter-proximal NusA and ρ peaks

Promoter-proximal RNAP peaks have been detected in E. coli and Drosophila, and are suggested to reflect RNAPs kinetically blocked early in elongation (for Drosophila) or possibly even prior to promoter escape (for E. coli; Core and Lis, 2008; Muse et al., 2007; Reppas et al., 2006; Wade and Struhl, 2008; Zeitlinger et al., 2007). Therefore, we asked whether promoter-proximal RNAP peaks were associated with NusA and ρ, which presumably requires promoter escape. We first calculated the traveling ratio (TR; the ratio of RNAP signal in the promoter-proximal peak to that within the TU; Reppas et al., 2006) for a set of genes with a 5′-σ70 peak and that were greater than 1 kb in length (to insure the peak and mid-gene signals were well separated; Fig. 6A). We then tested whether a NusA peak, ρ peak, or both occurred within 300 bp of the RNAP peak and binned the results based on TR (Fig. 6B). If the RNAP peaks reflect RNAPs poised prior to promoter escape, then the fraction of RNAP peaks with NusA or ρ co-peaks should decrease at low TR (because a low TR would indicate promoter-bound RNAP that should not recruit NusA or ρ, in contrast to ECs that can bind both). Instead, we observed little change in the frequency of NusA and ρ co-peaks at low TR.

Fig. 6
Frequency of ρ and NusA co-occurrence for RNAP peaks associated with genes

We also binned the frequency of NusA and ρ co-peaks based on gene expression level (Allen et al., 2003), to ask if a block to promoter escape correlates with low expression (as suggested previously by Reppas et al., 2006; Fig. 6C). No correlation was evident. Further, the frequency of co-peaks correlated to RNAP peak height (Fig. 6D), suggesting that the failure to detect NusA or ρ co-peaks for a fraction of RNAP peaks (~25%) is mostly explained by false negatives in the peak-calling algorithm, since the signal-to-noise ratio for RNAP is better than that for NusA or ρ. Taken together, these results suggest that promoter-proximal RNAP peaks reflect RNAPs that have escaped promoters, at which point signals for NusA and ρ become detectable.

To verify that RNAP peaks reflected premature termination rather than a block to promoter escape, we used quantitative RT-PCR to test representative sets of TUs that exhibited or lacked RNAP peaks (Figs. 4B,C) for a drop in RNA transcript levels. This is an imperfect test because RNAs generated by premature termination are more difficult than long mRNAs to quantify accurately and also may be unstable. Nonetheless, 6 of 8 TUs exhibiting RNAP peaks produced significantly more RNA near the 5′ end versus 0 of 4 for TUs lacking RNA peaks (Fig. S5; p<0.005; Student’s t-test). Thus, most RNAP peaks are associated with premature transcription termination. Reppas et al. (2006) raised the possibility that RNAP peaks might instead correspond to RNAPs poised prior to promoter escape in part because they found 300 σ70 peaks not associated with detectable mRNAs. Thus, we asked if these σ70 peaks exhibited NusA or ρ co-peaks. Of the 300 peaks, 20 correspond to highly expressed stable RNA genes; 138 of the remainder were associated with an RNAP peak (Table S6). Of these 138, 74 were within 300 bp of σ70 and RNAP peaks in our data. Of these 74, 45 (61%) were associated with a NusA peak; 49 (66%) were associated with a ρ peak; 33 (46%) were associated with both; and 13 (18%) were associated with neither (Fig. S6). As noted above, some NusA and ρ co-peaks for small RNAP peaks were probably missed. Nonetheless, a few RNAP peaks likely represent promoter-bound enzyme: of three examples specifically cited by Reppas et al. (2006), one (hepA) was associated with NusA and ρ, but two (deoB and yjiT) were associated with neither (data not shown).

ρ-dependent termination is not the primary cause of promoter-proximal RNAP peaks

The finding that promoter-proximal RNAP peaks correspond to RNAPs blocked early in elongation raised the possibility they result from transcriptional attenuation. Indeed, the Occapp profiles of genes regulated by attenuation resembled the aggregate profiles of genes associated with promoter-proximal RNAP peaks (Fig. S7). To ask if ρ, which also forms promoter-proximal peaks, could cause the RNAP peaks by ρ-dependent attenuation before a ribosome can bind and initiate translation, we examined the effect of the well-characterized ρ inhibitor, bicyclomycin (Supplemental Experimental Procedures). If the RNAP peaks were caused by ρ-dependent attenuation, they should be reduced when cells are treated with bicyclomycin. Instead, we observed little effect on the aggregate RNAP Occapp profiles of genes exhibiting promoter-proximal RNAP peaks (Fig. 6E), even though pronounced effects were evident on a gene known to be regulated by ρ-dependent attenuation (rho; Fig. 6G; Matsumoto et al., 1986). We also examined the effect of ρ inhibition on TR and observed little if any effect (Fig. 6F). Consistent with this result, there also is no preferential effect of even higher levels of bicyclomycin on expression of genes that exhibit low TRs (Cardinale et al., 2008; Fig. S8). Thus, ρ-dependent attenuation does not appear to be the principal cause of promoter-proximal RNAP peaks.


Our ChIP-chip study of the distributions of RNAP, σ70, NusA, NusG, and ρ on E. coli TUs reveals the patterns of trafficking for regulators most central to control of transcript elongation in bacteria, and has important implications for understanding the mechanisms underlying these patterns. σ70, NusA, NusG, and ρ are distributed relatively uniformly among most transcribing RNAP molecules with apparent relative affinities for elongating RNAP of NusA≈NusG>ρ>σ70. As RNAP moves away from a promoter, crosslinking of σ70 greatly decreases. ρ and NusA appear to associate with RNAP as σ70 association decreases, with ρ slightly preceding NusA, whereas NusG associates with elongating RNAP more slowly. As previously reported (Reppas et al., 2006), RNAP exhibits strong promoter-proximal peaks on many, but not all TUs. We find that these peaks correspond to ECs and that they do not result from ρ-dependent attenuation.

NusA, NusG, and ρ exhibit different patterns of EC association, but no TU-specific specialization

Our finding that NusA, NusG, and ρ are, to a first approximation, uniformly associated with ECs on most TUs suggests they act as general modulators of transcript elongation with about equal probability of altering responses of RNAP to intrinsic pause, arrest or termination sites, regardless of where these sites occur in the genome. Due to the limited resolution of ChIP-chip, this doesn’t preclude specific associations of regulators at intrinsic sites that affect only a minority of elongating RNAP molecules or at which events occur rapidly relative to movement of RNAP over the surrounding DNA sequences. The results do rule out the possibilities that NusA, NusG, or ρ associate with certain TUs or certain sites within TUs to the exclusion of other TUs or locations. Nonetheless, each regulator associates with ECs as they move away from promoters in a distinct, regulator-specific pattern that is similar on most TUs (Fig. 7).

Fig. 7
Model of transcription regulator trafficking during initiation to elongation transition

NusA exhibits negligible signal at promoters and associates with RNAP as σ70 association is lost, closely paralleling RNAP levels once RNAP moves away from a promoter (Figs. (Figs.22--4).4). NusA’s highest affinity contacts occur between the NusA CTD and the α-subunit CTD; additional contacts are made by NusA’s KH and S1 domains to the nascent RNA and by the NusA NTD to a second site on RNAP, which may include the β-subunit flap tip (Liu et al., 1996; Mah et al., 2000; Toulokhonov et al., 2001). At promoters, the α CTD binds to upstream DNA, either sequence-specifically at UP elements or non-specifically in association with σ70 (Estrem et al., 1999), and σ70 region 4 occupies the flap-tip until nascent RNA reaches 16-17 nt in length (Murakami et al., 2002; Nickels et al., 2006). Thus, NusA contacts are either not possible (to nascent RNA) or masked by DNA or σ70 until RNAP moves away from the promoter, at which point the association of NusA with the α CTD and nascent transcript likely tether NusA to the EC via interactions that are largely independent of EC position in a TU (Fig. 7).

Like NusA, NusG exhibits negligible signal at promoters, but unlike NusA appears to associate with RNAP in two phases. In the first phase, evident in aggregate Occapp profiles (Fig. 4), NusG increases association with RNAP rapidly to ~1 kb downstream from promoters. This first phase is distinct from NusA association both in the slower rise (NusA association appears to be complete by 300 bp into TUs) and in that NusG signal does not mirror the promoter-proximal RNAP peaks (Fig. 4C). In the second phase, NusG Occapp increases more slowly, resulting in the increased NusG/RNAP ratios for genes farther from promoters (Figs. 5D-E).

One explanation for the delayed association pattern of NusG could be competition with σ70 for its binding location on RNAP. NusG is suggested to bind RNAP via contacts to the clamp helices (Belogurov et al., 2007), which also make the tightest RNAP contact to σ70 (via σ70 region 2; Arthur and Burgess, 1998; Young et al., 2001). Although σ70 region 4 dissociates from the flap tip when 16-17 nt of RNA are synthesized, the σ70 region 2-clamp helices interaction can persist in the EC without steric conflict (Mooney et al., 2005). In this case, slow NusG association could reflect delayed dissociation of σ70 region 2. This would mean that σ70 dissociates from RNAP more slowly than reported by the ChIP-chip assay, which instead shows a sharp fall off in σ70 crosslinking immediately downstream from promoters (Fig. 4; Raffaelle et al., 2005; Reppas et al., 2006; Wade and Struhl, 2004; see below). Alternatively, σ70 may release rapidly and NusG binding could require long RNA transcripts since it has been suggested that NusG contains an RNA-binding activity (Steiner et al., 2002).

ρ associates with TUs closer to promoters than either NusA or NusG, and then appears to decrease somewhat in TU association farther from promoters, with an approximately uniform association relative to RNAP signal (Figs. (Figs.33--4).4). The location of the promoter-proximal ρ peak is consistent with the requirement of 80-100 nt for ρ effects on ECs (Lau and Roberts, 1985). Thus, ρ appears to bind as soon as the requisite nascent transcript becomes available, but perhaps fails to terminate transcription because NusG is not yet associated with RNAP. This early binding could position ρ to detect and subsequently terminate synthesis of the occasional mRNA on which translation fails. The strong ρ ChIP signal may be reduced once ribosomes load onto nascent RNA and prevent ρ from translocating close to RNAP (Fig. 7).

σ70 appears to associate with ECs stochastically

Our analysis of σ70 confirmed prior reports that the great majority of σ70 ChIP signal is lost as RNAP escapes the promoter (Raffaelle et al., 2005; Reppas et al., 2006; Wade and Struhl, 2004), but asymmetry of the σ70 aggregate Occapp peak suggests the signal is lost on average ~20 bp into TUs (Fig. S3). However, a low σ70 ChIP signal was present and was correlated with RNAP signal at the middle of TUs (Fig. 3B). This likely reflects σ70-EC interaction, although we cannot exclude other possibilities (e.g., that transcription increases non-specific binding of σ70-containing holoenzyme to DNA, for instance by removing nucleoid proteins from DNA). In any case, it is difficult to assess the extent of the interaction from the low σ70 ChIP signal because it may reflect far less efficient σ70 crosslinking to DNA than for promoter complexes (e.g., indirect σ70-RNAP and RNAP-DNA crosslinking in ECs rather than direct σ70-promoter DNA crosslinking). Our results are consistent with the view that σ70 breaks DNA contact when RNAP escapes a promoter after which σ70’s weakened contacts to RNAP allow its stochastic release (Mooney et al., 2005; Raffaelle et al., 2005; Shimamoto et al., 1986) but still support at least a weak equilibrium association with ECs and σ70 rebinding at promoter-like sequences encountered during elongation (Mooney et al., 2005; Mooney and Landick, 2003).

The mechanistic basis of promoter-proximal RNAP peaks

In principal, promoter-proximal RNAP peaks could reflect one of at least three mechanistically distinct types of blocks to transcription. RNAP could be trapped (1) prior to promoter escape (e.g., before strand opening or in abortive initiation); (2) early in elongation in a paused (or poised) state from which it can be released to productive elongation; or (3) by premature and presumably regulated transcription termination (transcriptional attenuation). Promoter-proximal RNAP peaks are common for human and Drosophila genes where they appear to be correlated with developmentally regulated rather than with housekeeping genes (ENCODE Project Consortium, 2004; Guenther et al., 2007; Muse et al., 2007; Zeitlinger et al., 2007). These peaks have been attributed to promoter-proximal pausing based on several criteria (Core and Lis, 2008; Muse et al., 2007; Zeitlinger et al., 2007). In S. cerevisiae, promoter-proximal peaks occur only in stationary phase and by unknown mechanism (Wade and Struhl, 2008). All three types of mechanisms are well characterized in E. coli: promoter-trapping (Laishram and Gowrishankar, 2007; Rosenthal et al., 2008), promoter-proximal pausing (Marr and Roberts, 2000; Hatoum and Roberts, 2008), and attenuation (Merino and Yanofsky, 2005).

Our findings establish that most promoter-proximal E. coli RNAP peaks correspond to ECs. First, the promoter-proximal RNAP peaks were offset in the direction of transcription by ~150 bp (Fig. 4). The transition from abortive to productive elongation, marked by release of σ70 from promoter contacts (or from RNAP contacts), occurs within the first 20 nt of transcript elongation (Chander et al., 2007; Revyakin et al., 2006). Known cases of σ70-stimulated pausing in vivo occur no later than +25 (Ring et al., 1996). Thus, the location of RNAP peaks at +150 is inconsistent with a block prior to promoter escape and EC formation. Second, NusA, which is thought to bind to ECs after release of σ70, and ρ, which requires >50 nt of RNA to bind, both appeared to be associated with RNAP in the promoter-proximal peaks.

Assuming that ChIP-chip captures a close-to-instantaneous snapshot of RNAP positions on DNA, we suggest that the promoter-proximal RNAP peaks reflect transcriptional attenuation caused by a mechanism other than ρ-dependent termination, rather than RNAP poised at promoters (Wade and Struhl, 2008). The position of these RNAP peaks is consistent with the typical position of transcription attenuators (Merino and Yanofsky, 2005) and strongly resembles ChIP-chip profiles of RNAP on TUs known to be subject to transcriptional attenuation (e.g., trp and pyrBI; Fig. S7). Promoter-proximal peaks in eukaryotes have been ascribed to paused ECs (Core and Lis, 2008; Muse et al., 2007; Zeitlinger et al., 2007). Although long elusive, transcription attenuation is now clearly shown to occur in eukaryotes (Steinmetz et al., 2006). Conclusive evidence that promoter-proximal halted RNAPs are actually paused rather than on a termination pathway exists only for a limited number of cases (e.g., Drosophila heat shock genes and bacteriophage λ PR; Adelman et al., 2005; Marr and Roberts, 2000). The regulation of early elongation by attenuation may prove to be more common in all organisms than has been appreciated.

Experimental Procedures

For additional information, see Supplemental Experimental Procedures.


E. coli K12 strains MG1655 and MG1655 HA3::nusG were used for all experiments. MG1655 HA3::nusG was constructed by gene replacement without selection to give a strain isogenic to MG1655 encoding three copies of the haemagglutinin (HA) epitope tag at the 5′ end of nusG. Monoclonal antibodies against σ70 (2G10), RNAP (anti-β’, NT73 or anti-β, NT63), and NusA (1NA1) were purchased from Neoclone (Madison, WI). The monoclonal 12CA5 anti-HA antibody (to target HA3::nusG) was purchased from Roche. The polyclonal antibody against NusG was generated by Proteintech (Chicago) and polyclonal antibody against ρ was a kind gift from Jeff Roberts (Cornell U.) After labeling, ChIP samples were hybridized to a custom microarray from Nimblegen (Madison, WI) that contains two copies of 187,204 Tm-matched ≥45mer oligonucleotides that tile the E. coli chromosome with an average of spacing of 24.5 bp.

Cell growth and ChIP-chip

Cells were grown in defined minimal medium (with 0.2% glucose) with vigorous shaking at 37 °C to mid-log (light scattering at 600 nm equivalent to 0.4 OD). Formaldehyde was added to 1% final and shaking was continued for 5 min before quenching with glycine. Cells were harvested, washed with PBS, and stored at -80 °C. Cells were sonicated and digested with micrococcal nuclease and RNase A before immunoprecipitation. The ChIP DNA sample was amplified by ligation-mediated PCR (Lee et al., 2006) to yield >4 μg of DNA, pooled with two other independent samples, and sent to Nimblegen where samples were labeled with Cy3 and Cy5 fluorescent dyes (one for the ChIP sample and one for a control input sample) and hybridized to a single microarray as a two-color experiment.

Supplementary Material




We thank C. Herring for construction of the HA3::nusG allele, K. Struhl for helpful discussions and sharing results prior to publication, and J. Grass for assistance with a control experiment. This work was supported by grants to A.Z.A. (USDA Hatch) and R. L. ( NIH GM38660).


  • Adelman K, Marr MT, Werner J, Saunders A, Ni Z, Andrulis ED, Lis JT. Efficient release from promoter-proximal stall sites requires transcript cleavage factor TFIIS. Mol Cell. 2005;17:103–112. [PubMed]
  • Allen TE, Herrgard MJ, Liu M, Qiu Y, Glasner JD, Blattner FR, Palsson BO. Genome-scale analysis of the uses of the Escherichia coli genome: model-driven analysis of heterogeneous data sets. J Bacteriol. 2003;185:6392–6399. [PMC free article] [PubMed]
  • Arthur TM, Burgess RR. Localization of a sigma70 binding site on the N terminus of the Escherichia coli RNA polymerase beta’ subunit. J Biol Chem. 1998;273:31381–31387. [PubMed]
  • Artsimovitch I, Landick R. Pausing by bacterial RNA polymerase is mediated by mechanistically distinct classes of signals. Proc Natl Acad Sci U S A. 2000;97:7090–7095. [PMC free article] [PubMed]
  • Artsimovitch I, Landick R. The transcriptional regulator RfaH stimulates RNA chain synthesis after recruitment to elongation complexes by the exposed nontemplate DNA strand. Cell. 2002;109:193–203. [PubMed]
  • Bar-Nahum G, Nudler E. Isolation and characterization of sigma(70)-retaining transcription elongation complexes from Escherichia coli. Cell. 2001;106:443–451. [PubMed]
  • Belogurov GA, Vassylyeva MN, Svetlov V, Klyuyev S, Grishin NV, Vassylyev DG, Artsimovitch I. Structural basis for converting a general transcription factor into an operon-specific virulence regulator. Mol Cell. 2007;26:117–129. [PMC free article] [PubMed]
  • Burns CM, Richardson LV, Richardson JP. Combinatorial effects of NusA and NusG on transcription elongation and Rho-dependent termination in Escherichia coli. J Mol Biol. 1998;278:307–316. [PubMed]
  • Cardinale CJ, Washburn RS, Tadigotla VR, Brown LM, Gottesman ME, Nudler E. Termination factor Rho and its cofactors NusA and NusG silence foreign DNA in E. coli. Science. 2008;320:935–938. [PMC free article] [PubMed]
  • Chander M, Austin KM, Aye-Han NN, Sircar P, Hsu LM. An alternate mechanism of abortive release marked by the formation of very long abortive transcripts. Biochemistry. 2007;46:12687–12699. [PubMed]
  • Core LJ, Lis JT. Transcription regulation through promoter-proximal pausing of RNA polymerase II. Science. 2008;319:1791–1792. [PMC free article] [PubMed]
  • Defez R, De Felice M. Cryptic operon for beta-glucoside metabolism in Escherichia coli K12: genetic evidence for a regulatory protein. Genetics. 1981;97:11–25. [PMC free article] [PubMed]
  • deHaseth PL, Lohman TM, Burgess RR, Record MT., Jr. Nonspecific interactions of Escherichia coli RNA polymerase with native and denatured DNA: differences in the binding behavior of core and holoenzyme. Biochemistry. 1978;17:1612–1622. [PubMed]
  • ENCODE Project Consortium The ENCODE (ENCyclopedia Of DNA Elements) Project. Science. 2004;306:636–640. [PubMed]
  • Estrem ST, Ross W, Gaal T, Chen ZW, Niu W, Ebright RH, Gourse RL. Bacterial promoter architecture: subsite structure of UP elements and interactions with the carboxy-terminal domain of the RNA polymerase alpha subunit. Genes Dev. 1999;13:2134–2147. [PMC free article] [PubMed]
  • Farnham PJ, Greenblatt J, Platt T. Effects of NusA protein on transcription termination of the tryptophan operon of Escherichia coli. Cell. 1982;29:945–951. [PubMed]
  • Grainger DC, Hurd D, Harrison M, Holdstock J, Busby SJ. Studies of the distribution of Escherichia coli cAMP-receptor protein and RNA polymerase along the E. coli chromosome. Proc Natl Acad Sci U S A. 2005;102:17693–17698. [PMC free article] [PubMed]
  • Greenblatt J, Li J. Interaction of the sigma factor and the nusA gene protein of E. coli with RNA polymerase in the initiation-termination cycle of transcription. Cell. 1981;24:421–428. [PubMed]
  • Greenblatt J, McLimont M, Hanly S. Termination of transcription by nusA gene protein of Escherichia coli. Nature. 1981;292:215–220. [PubMed]
  • Grigorova IL, Phleger NJ, Mutalik VK, Gross CA. Insights into transcriptional regulation and sigma competition from an equilibrium model of RNA polymerase binding to DNA. Proc Natl Acad Sci U S A. 2006;103:5332–5337. [PMC free article] [PubMed]
  • Guenther MG, Levine SS, Boyer LA, Jaenisch R, Young RA. A chromatin landmark and transcription initiation at most promoters in human cells. Cell. 2007;130:77–88. [PMC free article] [PubMed]
  • Hatoum A, Roberts J. Prevalence of RNA polymerase stalling at Escherichia coli promoters after open complex formation. Mol Microbiol. 2008;68:17–28. [PMC free article] [PubMed]
  • Herring CD, Raffaelle M, Allen TE, Kanin EI, Landick R, Ansari AZ, Palsson BO. Immobilization of Escherichia coli RNA polymerase and location of binding sites by use of chromatin immunoprecipitation and microarrays. J Bacteriol. 2005;187:6166–6174. [PMC free article] [PubMed]
  • Kapanidis AN, Margeat E, Laurence TA, Doose S, Ho SO, Mukhopadhyay J, Kortkhonjia E, Mekler V, Ebright RH, Weiss S. Retention of transcription initiation factor sigma70 in transcription elongation: single-molecule analysis. Mol Cell. 2005;20:347–356. [PubMed]
  • Kassavetis GA, Chamberlin MJ. Pausing and termination of transcription within the early region of bacteriophage T7 DNA in vitro. J Biol Chem. 1981;256:2777–2786. [PubMed]
  • Kuo MH, Allis CD. In vivo cross-linking and immunoprecipitation for studying dynamic Protein:DNA associations in a chromatin environment. Methods. 1999;19:425–433. [PubMed]
  • Laishram RS, Gowrishankar J. Environmental regulation operating at the promoter clearance step of bacterial transcription. Genes Dev. 2007;21:1258–1272. [PMC free article] [PubMed]
  • Lau LF, Roberts JW. Rho-dependent transcription termination at lambda R1 requires upstream sequences. J Biol Chem. 1985;260:574–584. [PubMed]
  • Lee TI, Johnstone SE, Young RA. Chromatin immunoprecipitation and microarray-based analysis of protein location. Nat Protoc. 2006;1:729–748. [PMC free article] [PubMed]
  • Li J, Horwitz R, McCracken S, Greenblatt J. NusG, a new Escherichia coli elongation factor involved in transcriptional antitermination by the N protein of phage lambda. J Biol Chem. 1992;267:6012–6019. [PubMed]
  • Li J, Mason SW, Greenblatt J. Elongation factor NusG interacts with termination factor rho to regulate termination and antitermination of transcription. Genes Dev. 1993;7:161–172. [PubMed]
  • Linn T, Greenblatt J. The NusA and NusG proteins of Escherichia coli increase the in vitro readthrough frequency of a transcriptional attenuator preceding the gene for the beta subunit of RNA polymerase. J Biol Chem. 1992;267:1449–1454. [PubMed]
  • Liu K, Zhang Y, Severinov K, Das A, Hanna MM. Role of Escherichia coli RNA polymerase alpha subunit in modulation of pausing, termination and anti-termination by the transcription elongation factor NusA. EMBO J. 1996;15:150–161. [PMC free article] [PubMed]
  • Mah TF, Kuznedelov K, Mushegian A, Severinov K, Greenblatt J. The alpha subunit of E. coli RNA polymerase activates RNA binding by NusA. Genes Dev. 2000;14:2664–2675. [PMC free article] [PubMed]
  • Marr MT, Roberts JW. Function of transcription cleavage factors GreA and GreB at a regulatory pause site. Mol Cell. 2000;6:1275–1285. [PubMed]
  • Mason SW, Li J, Greenblatt J. Host factor requirements for processive antitermination of transcription and suppression of pausing by the N protein of bacteriophage lambda. J Biol Chem. 1992;267:19418–19426. [PubMed]
  • Matsumoto Y, Shigesada K, Hirano M, Imai M. Autogenous regulation of the gene for transcription termination factor rho in Escherichia coli: localization and function of its attenuators. J Bacteriol. 1986;166:945–958. [PMC free article] [PubMed]
  • Merino E, Yanofsky C. Transcription attenuation: a highly conserved regulatory strategy used by bacteria. Trends Genet. 2005;21:260–264. [PubMed]
  • Mooney RA, Darst SA, Landick R. Sigma and RNA polymerase: an on-again, off-again relationship? Mol Cell. 2005;20:335–345. [PubMed]
  • Mooney RA, Landick R. Tethering σ70 to RNA polymerase reveals high in vivo activity of σ factors and σ70-dependent pausing at promoter-distal locations. Genes Dev. 2003;17:2839–2851. [PMC free article] [PubMed]
  • Mukhopadhyay J, Kapanidis AN, Mekler V, Kortkhonjia E, Ebright YW, Ebright RH. Translocation of sigma(70) with RNA polymerase during transcription: fluorescence resonance energy transfer assay for movement relative to DNA. Cell. 2001;106:453–463. [PubMed]
  • Murakami KS, Masuda S, Campbell EA, Muzzin O, Darst S. Structural basis of transcription Initiation: an RNA polymerase holoenzyme/DNA complex. Science. 2002;296:1285–1290. [PubMed]
  • Muse GW, Gilchrist DA, Nechaev S, Shah R, Parker JS, Grissom SF, Zeitlinger J, Adelman K. RNA polymerase is poised for activation across the genome. Nat Genet. 2007;39:1507–1511. [PMC free article] [PubMed]
  • Nickels BE, Roberts CW, Roberts JW, Hochschild A. RNA-mediated destabilization of the sigma(70) region 4/beta flap interaction facilitates engagement of RNA polymerase by the Q antiterminator. Mol Cell. 2006;24:457–468. [PMC free article] [PubMed]
  • Raffaelle M, Kanin EI, Vogt J, Burgess RR, Ansari AZ. Holoenzyme switching and stochastic release of sigma factors from RNA polymerase in vivo. Mol Cell. 2005;20:357–366. [PubMed]
  • Reppas NB, Wade JT, Church GM, Struhl K. The Transition between transcriptional initiation and elongation in E. coli Is highly variable and often rate limiting. Mol Cell. 2006;24:747–757. [PubMed]
  • Revyakin A, Liu C, Ebright RH, Strick TR. Abortive initiation and productive initiation by RNA polymerase involve DNA scrunching. Science. 2006;314:1139–1143. [PMC free article] [PubMed]
  • Richardson J. Rho-dependent termination and ATPases in transcript termination. Biochim Biophys Acta. 2002;1577:251. [PubMed]
  • Ring B, Yarnell W, Roberts J. Function of E. coli RNA polymerase σ factor σ70 in promoter-proximal pausing. Cell. 1996;86:485–493. [PubMed]
  • Rosenthal AZ, Kim Y, Gralla JD. Poising of Escherichia coli RNA polymerase and its release from the sigma 38 C-terminal tail for osmY transcription. J Mol Biol. 2008;376:938–949. [PMC free article] [PubMed]
  • Selinger DW, Cheung KJ, Mei R, Johansson EM, Richmond CS, Blattner FR, Lockhart DJ, Church GM. RNA expression analysis using a 30 base pair resolution Escherichia coli genome array. Nat Biotechnol. 2000;18:1262–1268. [PubMed]
  • Shankar S, Hatoum A, Roberts JW. A transcription antiterminator constructs a NusA-dependent shield to the emerging transcript. Mol Cell. 2007;27:914–927. [PMC free article] [PubMed]
  • Shimamoto N, Kamigochi T, Utiyama H. Release of the sigma subunit of Escherichia coli DNA-dependent RNA polymerase depends mainly on time elapsed after the start of initiation, not on length of product RNA. J Biol Chem. 1986;261:11859–11865. [PubMed]
  • Solomon MJ, Larsen PL, Varshavsky A. Mapping protein-DNA interactions in vivo with formaldehyde: evidence that histone H4 is retained on a highly transcribed gene. Cell. 1988;53:937–947. [PubMed]
  • Steiner T, Kaiser JT, Marinkovic S, Huber R, Wahl MC. Crystal structures of transcription factor NusG in light of its nucleic acid- and protein-binding activities. Embo J. 2002;21:4641–4653. [PMC free article] [PubMed]
  • Steinmetz EJ, Warren CL, Kuehner JN, Panbehi B, Ansari AZ, Brow DA. Genome-wide distribution of yeast RNA polymerase II and its control by Sen1 helicase. Mol Cell. 2006;24:735–746. [PubMed]
  • Sullivan SL, Gottesman ME. Requirement for E. coli NusG protein in factor-dependent transcription termination. Cell. 1992;68:989–994. [PubMed]
  • Torres M, Condon C, Balada JM, Squires C, Squires CL. Ribosomal protein S4 is a transcription factor with properties remarkably similar to NusA, a protein involved in both non-ribosomal and ribosomal RNA antitermination. Embo J. 2001;20:3811–3820. [PMC free article] [PubMed]
  • Toulokhonov I, Artsimovitch I, Landick R. Allosteric control of RNA polymerase by a site that contacts nascent RNA hairpins. Science. 2001;292:730–733. [PubMed]
  • von Hippel PH, Revzin A, Gross CA, Wang AC. Non-specific DNA binding of genome regulating proteins as a biological control mechanism: I. The lac operon: equilibrium aspects. Proc Natl Acad Sci U S A. 1974;71:4808–4812. [PMC free article] [PubMed]
  • Wade JT, Struhl K. Association of RNA polymerase with transcribed regions in Escherichia coli. Proc Natl Acad Sci U S A. 2004;101:17777–17782. Epub 12004 Dec 17713. [PMC free article] [PubMed]
  • Wade JT, Struhl K. The transition from transcriptional initiation to elongation. Curr Opin Genet Dev. 2008;18:130–136. [PMC free article] [PubMed]
  • Wade JT, Struhl K, Busby SJ, Grainger DC. Genomic analysis of protein-DNA interactions in bacteria: insights into transcription and chromosome organization. Mol Microbiol. 2007;65:21–26. [PubMed]
  • Yakhnin AV, Babitzke P. NusA-stimulated RNA polymerase pausing and termination participates in the Bacillus subtilis trp operon attenuation mechanism invitro. Proc Natl Acad Sci U S A. 2002;99:11067–11072. [PMC free article] [PubMed]
  • Young BA, Anthony LC, Gruber TM, Arthur TM, Heyduk E, Lu CZ, Sharp MM, Heyduk T, Burgess RR, Gross CA. A coiled-coil from the RNA polymerase beta’ subunit allosterically induces selective nontemplate strand binding by sigma(70) Cell. 2001;105:935–944. [PubMed]
  • Zeitlinger J, Stark A, Kellis M, Hong JW, Nechaev S, Adelman K, Levine M, Young RA. RNA polymerase stalling at developmental control genes in the Drosophila melanogaster embryo. Nat Genet. 2007;39:1512–1516. [PMC free article] [PubMed]
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...