Search tips
Search criteria 


Logo of jbacterPermissionsJournals.ASM.orgJournalJB ArticleJournal InfoAuthorsReviewers
J Bacteriol. 2009 December; 191(24): 7436–7446.
Published online 2009 October 9. doi:  10.1128/JB.00412-09
PMCID: PMC2786607

Growth Phase and (p)ppGpp Control of IraD, a Regulator of RpoS Stability, in Escherichia coli[down-pointing small open triangle]


The antiadaptor protein IraD inhibits the proteolysis of the alternative sigma factor, RpoS, which promotes the synthesis of >100 genes during the general stress response and during stationary phase. Our previous results showed that IraD determines RpoS steady-state levels during exponential growth and mediates its stabilization after DNA damage. In this study, we show by promoter fusions that iraD was upregulated during the transition from exponential growth to stationary phase. The levels of RpoS likewise rose during this transition in a partially IraD-dependent manner. The expression of iraD was under the control of ppGpp. The expression of iraD required RelA and SpoT (p)ppGpp synthetase activities and was dramatically induced by a “stringent” allele of RNA polymerase, culminating in elevated levels of RpoS. Surprisingly, DksA, normally required for transcriptional effects of the stringent response, repressed iraD expression, suggesting that DksA can exert regulatory effects independent of and opposing those of (p)ppGpp. Northern blot analysis and 5′ rapid amplification of cDNA ends revealed two transcripts for iraD in wild-type strains; the smaller was regulated positively by RelA during growth; the larger transcript was induced specifically upon transition to stationary phase and was RelA SpoT dependent. A reporter fusion to the distal promoter indicated that it accounts for growth-phase regulation and DNA damage inducibility. DNA damage inducibility occurred in strains unable to synthesize (p)ppGpp, indicating an additional mode of regulation. Our results suggest that the induction of RpoS during transition to stationary phase and by (p)ppGpp occurs at least partially through IraD.

Bacteria such as Escherichia coli control gene expression in response to environmental signals. One component of this control is the use of alternative sigma factors that change promoter recognition by RNA polymerase and therefore the repertoire of expressed genes. The transcription factor σS, or RpoS, controls the synthesis of more than 100 genes, many of which are induced during the stationary phase of growth (15, 16, 24, 33, 45). RpoS levels respond sensitively to growth rate (42) and σS accumulates during the transition from proliferative growth to quiescence (17, 26, 41). Even during exponential growth, σS controls gene expression in response to a wide variety of environmental insults, comprising the so-called “general stress response” (reviewed in references 19, 21, and 27).

RpoS is regulated at every conceivable level: transcriptional, translational, protein stability, and activity (reviewed in references 20 and 23), and the integration of signals that impinge on RpoS is complex. Posttranslational control appears to underlie much of the regulation of σS in response to stress during growth. During cell proliferation, σS is synthesized but rapidly degraded, with a half-life of 2 to 3 min, but RpoS becomes stabilized upon entry into stationary phase (26, 40). The ClpXP protease degrades RpoS but does not directly recognize the protein; degradation requires the adaptor molecule RssB (5, 30, 31, 38, 47), which binds RpoS directly (48). RssB levels are limiting in the cell (39) and can be titrated by increased RpoS synthesis. RssB is a two-component response regulator protein, but the consequence of phosphorylation of the aspartate in the predicted receiver domain (8) on RpoS stability is negligible (36).

In the presence of stress such as starvation or DNA damage, RpoS levels are induced by the synthesis of specific small antiadaptor molecules that bind RssB and prevent its action as an adaptor for RpoS proteolysis. The first of these to be discovered, IraP, is induced by phosphate starvation (11); IraD, the subject of the present study, is induced by DNA damage (9, 29); IraM is induced by magnesium starvation (9). Accumulation of any of these proteins leads to RpoS stabilization in vivo and protection from proteolysis in vitro (9, 11). The antiadaptor mechanism therefore allows the induction of situation-specific factors that can rapidly and transiently affect RpoS levels.

The induction of the σS regulon has been correlated with a second signal, the production of (p)ppGpp during the stringent response (reviewed in references 13 and 37). Levels of ppGpp (the more predominant of the two signaling molecules) and RpoS rise dramatically at the late exponential phase of growth (13, 17, 26, 41). (p)ppGpp is synthesized by two enzymes: RelA and SpoT (46). SpoT has dual functions and possesses the only known ppGpp hydrolase activity, regulated apart from its synthetase function (22, 32).

The best-studied (p)ppGpp response is the RelA-dependent stringent response to amino acid starvation mediated through effects of ppGpp on RNA polymerase activity. In the presence of uncharged tRNA, RelA is activated to synthesize (p)ppGpp (7). RNA polymerase bound by ppGpp has altered properties with respect to open complex formation, resulting in downregulation of stable RNA synthesis and upregulation of certain biosynthetic operons. At rRNA promoters such as the well-studied rrnB P1 promoter, ppGpp disfavors initiation by preventing stable open complex formation (3, 4). Upregulation of biosynthetic promoters may be indirect, by reallocation of RNAP newly freed from stable RNA synthesis, but ppGpp can also directly activate transcription in vitro, in the presence of the small transcription factor DksA (35). In vivo, DksA is required for the full effect of (p)ppGpp on both negatively and positively regulated promoters (34, 35). However, DksA has effects on phenotypes such as adhesion, motility and chemotaxis in relA spoT double mutant strains [so-called “(p)ppGpp0” strains] unable to synthesize (p)ppGpp (1, 28), suggesting that DksA can function independently of (p)ppGpp. Microarray and fusion experiments indicated that DksA and (p)ppGpp have opposing effects on a subset of genes primarily involved in fimbriae regulation, chemotaxis and motility (1, 2).

The connection between the stringent response and RpoS is complex. (p)ppGpp has been implicated in transcriptional (25), translational (12), and posttranslational (10) control of RpoS. Perhaps by a combination of these mechanisms, RpoS accumulates during the transition from exponential growth to stationary phase, and this is greatly diminished in (p)ppGpp0 strains (17). Artificial induction of high levels of (p)ppGpp leads to elevated RpoS in the cell (12, 17). At least one level of control impinges on the antiadaptor mechanism of RpoS regulation: induction of iraP during phosphate starvation and stabilization of RpoS requires ppGpp synthesis capacity (10).

In the present study we examine the effects of the stringent response on the IraD antiadaptor. We identified iraD in a mutant screen for sensitivity to a variety of DNA damaging agents (29). In the presence of DNA damage such as oxidative stress or the replication inhibitor, azidothymidine (AZT), SOS-independent induction of iraD leads to RpoS stabilization (29). Like IraP, IraD directly binds RssB and prevents proteolysis of RpoS by ClpXP (9). Unlike IraP, which has effects only on RpoS after starvation, IraD appears to promote higher steady-state levels of RpoS, even in the absence of stress (29). We show here that iraD is growth phase regulated and that (p)ppGpp accumulation is necessary for full iraD induction during the transition to stationary phase. In strains containing “stringent” alleles of RNA polymerase that mimic the transcriptional effects after (p)ppGpp accumulation (49), iraD is constitutively induced. In such strains, RpoS levels are also induced, in a partially IraD-dependent manner, suggesting that IraD is a critical regulator of RpoS under conditions when ppGpp accumulates, such as during the stringent response and during transition to stationary phase.


Strains, growth conditions.

All strains used in the present study are listed in Table Table11 and are isogenic with MG1655 (F rph-1) (6). Cultures were grown at 37°C in Luria-Bertani (LB) medium supplemented with appropriate antibiotics kanamycin (20 μg/ml), chloramphenicol (15 μg/ml), ampicillin (100 μg/ml), and/or tetracycline (10 μg/ml). Strains were constructed by P1 transductions and phage lysates grown in LCG (LB medium supplemented with 1% glucose and 2 mM calcium chloride). Strain 13119 [btuB::Tn10 spoT(E219Q)] was constructed with P1 donor CF11608 (provided by D. Jin [18]) crossed with MG1655, selecting tetracycline resistance; the presence of the spoT allele was confirmed by PCR, followed by sequence analysis. relAΔ::FRT and dksAΔ::FRT cat were both constructed by PCR amplification of the cat gene from pKD4 with Taq DNA polymerase, using primers carrying 40-nucleotide (nt) homology to the open reading frame (ORF) of each gene as described previously (14). After purification and DpnI digestion of the PCR fragment, it was transformed into MG1655 harboring recombination facilitator plasmid pKD46, selecting chloramphenicol resistance. pKD46 was subsequently cured by growth at 42°C.

Strains and plasmids used in this study

Plasmid constructions.

Plasmid iraD::luxCDABE (29) was constructed as previously described. Plasmids P1-iraD600/375::luxCDABE and P2-iraD262/1::luxCDABE were constructed from the primer pairs 5′-GGG GAC AAG TTT GTA CAA AAA AGC AGG CTT CGA AGG AGA TAG AAC CGT AAA CAA ATG ACA TGC ATG TTT CT-3′/5′-GGG GAC CAC TTT GTA CAA GAA AGC TGG GTC AAT TGA TAA CAT ATC AAA TAC ATC AAA AAT-3′ and 5′-GGG GAC AAG TTT GTA CAA AAA AGC AGG CTT CGA AGG AGA TAG AAC CCT ATT CGG AAT ATT CAG AAT ATA CTA GCG G-3′/5′-GGG GAC CAC TTT GTA CAA GAA AGC TGG GTC TTT GCG CAC TCC TGA CGT TTA GCA A-3′, respectively, as previously described (29).

Transcriptional start site (TSS) mapping.

The 5′ ends of iraD mRNA were mapped by using 5′ RACE (rapid amplification of cDNA ends) as previously described (43) with the following modifications. cDNA was prepared by incubation of 2 μg of total RNA (isolated as described below) with 200 U of Moloney murine leukemia virus reverse transcriptase (NEB), 2 mM deoxynucleoside triphosphate, 40 μM iraD ORF specific primer (5′-TTA GCT GAC ATT CTC CAG CGT CGC ACT GCG-3′), and 1× Moloney murine leukemia virus RT buffer at 42°C for 1 h, followed by treatment with 10 μg of RNase A (Sigma)/ml and 5 U of RNase H (NEB) at 37°C for 30 min. cDNA was purified by using a QIAquick PCR purification kit (Qiagen). After ligation of cDNA with an anchor oligonucleotide (5′-TTT AGT GAG GGT TAA TAA GCG GCC GCG TCG TGA CTG GGA GCG C-3′), touchdown PCR was performed by using 1 U of Phusion polymerase (Finnzyme), 12 ng of purified anchored-cDNA, 0.5 μM anchor, and iraD ORF specific primers (5′-CTC CCA GTC ACG ACG CGG CCG C-3′ and 5′-CAA ACG TTA GCG GTT CAT CG-3′). The PCR cycling conditions were as follows: 98°C for 3 min (1 cycle); 98°C for 30 s to 72°C (at 1°C/cycle) for 1 min (15 cycles); 72°C for 1.5 min (1 cycle); 98°C for 30 s, 58°C for 1 min, and 72°C for 1.5 min (20 cycles); 72°C for 5 min (1 cycle); followed finally by holding at 4°C. PCR products were resolved on 1.5% agarose gels and bands were excised by using a QIAEX II gel extraction kit (Qiagen) and sequenced using the iraD ORF and anchor-specific primers (5′-TTC TTT CAA ATT AAC CTG CAA CGC C-3′ and 5′-TAA AAA GAG TGA GGA GAT CGC-3′, respectively).

RpoS levels and stability experiments.

RpoS steady-state levels during growth were determined by Western blot analysis (29) from cultures harvested at the indicated phase of growth. The stability of RpoS was determined after the addition of chloramphenicol (200 μg/ml) to exponential-phase LB cultures harvested at the time points indicated. Protein was precipitated after a 30-min incubation on ice in 20% ice-cold trichloroacetic acid (TCA), and the pellet was resuspended in sodium dodecyl sulfate buffer. RpoS Western blots were performed as described previously (29).

Lux reporter assays.

Luciferase reporter assays were performed as described previously (29, 44). For expression measurements throughout growth, cultures were inoculated and grown overnight at 37°C in LB medium containing only ampicillin (100 μg/ml). After a 14-h incubation period, 25 ml of the same medium in the absence of antibiotics was inoculated with cells to an optical density at 600 nm (OD600) of 0.05. Time points (OD600 and counts per minute) were determined every 20 min after a 90-min initial incubation. All cultures were grown with aeration throughout the time course, and time points were taken at 37°C without the introduction of variability in temperature. In AZT induction experiments, cultures were grown as described above and treated with 1 μg of AZT/ml for 40 min before measurements were taken. For rpoB* experiments, time points were taken every 30 min after an initial 90-min incubation in 40 ml of LB medium. Luminescence from strains containing the reporter vector (with no inserted promoter region) was measured and did not significantly contribute to the reported data.

RNA purification and Northern blot analysis.

Cells were grown in LB as described above, and samples for exponential phase were taken at an optical density of 0.7 (± 0.1) and for stationary phase at 1.6 (± 0.1). RNA was isolated by using the RiboPure -Bacteria kit from Ambion according to the manufacturer's instructions. Concentrations of RNA samples were determined by measuring the OD260. Analysis of iraD mRNA by Northern blots was performed with 7.5 μg of total RNA per sample, loaded equally on a 1% agarose-formaldehyde denaturing gel, and ran in MOPS buffer. Samples were transferred onto a positively charged membrane (Amersham Hybond N+), cross-linked by using UV, and transcripts were detected by using a specific, 400-nt DNA probe complementary to iraD's ORF in ExpressHyb Hybridization Solution (Clontech). RNA integrity and equal loading of each sample lane was monitored via methylene blue staining of the membranes prior to hybridization. The iraD specific probe was constructed from the PCR product generated by a forward primer (5′-CAA TCA CTT CAG GCT GTT TTA CC-3′) and a reverse primer (5′-TTA GCT GAC ATT CTC CAG CGT CGC ACT GCG-3′). The PCR product was labeled according to manufacturer's instructions using ATP-αP32 and the Prime-It II Random Primer Labeling Kit (Stratagene). The sizes of the apparent iraD transcripts were estimated by using RNA Century-Plus Markers (Ambion). The graphs below the Northern blots show quantification of each transcript in that condition relative to 16S rRNA as quantitated by phosphorImage analysis (GE Storm 860, with ImageQuant TL software), in the strain indicated. The standard deviation in the quantifications (shown with error bars) was determined by the product of the number of pixels and standard deviation of the pixel intensity as calculated by Adobe Photoshop.


Two TSSs for iraD.

To define the promoters responsible for iraD expression, we performed 5′ RACE followed by sequence analysis of the resulting PCR product, both during exponential growth and during stationary phase (Fig. (Fig.1A).1A). Sequence analysis of DNA generated from cells in exponential and/or stationary phase revealed two mRNA start sites; one at −417 (referred to as P1) and a second at −137 (referred to as P2), relative to the ATG start codon of the iraD ORF (Fig. (Fig.1A).1A). A reasonable match to −10 and −35 consensus promoter sequences were found immediately upstream of both start sites (Fig. (Fig.1B).1B). Interestingly, the ~800-nt transcript generated from the distal promoter (P1) was present during entry into stationary phase but not in exponential phase, whereas the ~500-nt transcript generated from the proximal promoter (P2) is present at both growth phases. There is no known protein coding region in this upstream region of iraD; the largest predicted ORF could encode a 47-amino-acid peptide.

FIG. 1.
Mapping of the iraD TSSs and correlation of iraD expression with growth phase and RpoS levels. (A) Agarose gel of 5′ RACE PCR products generated from MG1655 cells in either logarithmic or stationary phase as indicated. Sequencing of the PCR products ...

iraD expression is growth phase dependent.

We examined iraD expression indirectly through fusion of 600 nt of its upstream intergenic region, including both promoter regions, to a luciferase operon reporter (29, 44) (iraD::luxCDABE) in a simple growth experiment. As apparent in Fig. Fig.1C,1C, iraD was expressed very highly in late exponential phase of growth, as determined from the optical density of the culture over time, and maintained at high levels during stationary phase. Expression reached a minimum as cells entered the middle exponential phase of growth.

The two iraD promoter regions show differences in their expression pattern.

To dissect the contribution of each promoter site to the expression pattern of the full-length promoter fusion construct, we constructed two new luciferase reporters containing either the distal P1 or the proximal P2 TSSs. In a growth experiment, we found that the reporter fused to the region containing the distal P1 (spanning −600 to −375 relative to the start of the ORF) behaved similarly to the full-length promoter construct with a strong induction during the transition to stationary phase, although the expression levels were somewhat higher in exponential growth than the full-length promoter fusion. The reporter construct containing P2 (spanning −262 to −1 relative to the start of ORF) did not behave like the full-length fusion reporter, and showed expression levels that were low and steady (Fig. (Fig.1C),1C), equivalent to that for the full-length fusion at its lowest point. These results suggest that the full-length promoter fusion construct may report the sum of both transcripts during exponential growth, with P1 predominating during the late stages of growth.

RpoS levels during normal growth are determined by IraD.

Previously, we showed that IraD is involved in regulating RpoS stability in response to DNA damage (29). Based on the expression pattern of iraD, here we sought to determine whether IraD is involved in regulating RpoS levels in the late logarithmic phase of growth. Western blot analysis showed that RpoS levels begin to rise during the transition from exponential growth to stationary phase (Fig. (Fig.1D)1D) as has been reported previously (12, 17). In mutants deficient in IraD, steady-state levels of RpoS were reduced by ca. 60% at mid to late exponential phase during growth, suggesting that IraD is an important factor promoting accumulation of RpoS as cells begin to enter stationary phase. Although at a much lower level, RpoS still accumulates in a similar time frame as in a wild-type background in the iraD mutant, indicating there are yet other factors influencing the timing of its accumulation.

iraD expression is positively influenced by both RelA and SpoT.

Because (p)ppGpp levels rise dramatically as cells enter the late exponential phase of growth, we examined the impact of the known (p)ppGpp synthetases on the expression of iraD. Using the iraD::lux reporter constructs, we looked at the expression profile of iraD in either a relA deletion mutant or in spoT(E319Q), a mutant specifically defective for (p)ppGpp synthetase activity (18). (We were unable to look at a spoT-null mutant since the ppGpp hydrolase activity of the protein is essential for survival for strains that can produce (p)ppGpp). We found that the expression of the gene was reduced in both the relA and spoT(E319Q) mutants strains throughout fast growth in the mid-exponential phase (Fig. (Fig.2A),2A), although induction in late exponential phase was still evident. After overnight growth into the stationary phase, both mutants seemed to reach wild-type levels of iraD expression, suggesting that either iraD expression is independent of (p)ppGpp altogether during this time or that one of the two enzymes can compensate for the absence of the other (Table (Table2).2). Supporting the latter notion, a relA spoT double-knockout mutant resulted in even lower levels of expression from the iraD::luxCDABE reporter compared to that of either single knockout in early exponential phase (Fig. (Fig.2B)2B) and in stationary phase, culminating in fivefold lower levels than in wild-type strains and single relA or spoT(E319E) mutants (Table (Table2).2). A fusion of the recA promoter to the same luciferase construct did not show any difference in expression in these (p)ppGpp-deficient backgrounds, suggesting that this is a specific effect on iraD and not the technique or the reporter fusion (data not shown). These results suggest that ppGpp production is one factor required for full induction of iraD expression during the transition to stationary phase of growth. We note, however, that in the absence of (p)ppGpp production, iraD induction does indeed occur as cells enter stationary phase, indicating an additional mode of growth regulation. This is quite similar to that observed for RpoS itself: in (p)ppGpp0 strains, RpoS accumulation is delayed and modestly reduced, but not abolished, during the transition to stationary phase (12, 17).

FIG. 2.
iraD::luxCDABE expression in strains affected for (p)ppGpp synthesis. Each data point is an average of four independent determinations. The variability is shown with error bars in both graphs. RLU, relative luminescence units (bioluminescence counts per ...
Late-stationary-phase expression of iraD::luxDCABE in strains affecting the stringent response

To determine stringent regulation at each of the two promoter regions of iraD, we introduced the reporter constructs containing either the P1 or the P2 promoter regions into the various mutant backgrounds, and assayed their expression throughout growth. The luciferase reporter constructs showed that the distal segment of the iraD upstream region containing the P1 start site was positively regulated by (p)ppGpp, and showed lower expression in the late logarithmic phase of growth in the relA mutant and in the relA spoT double-mutant backgrounds (Fig. (Fig.2C).2C). In the spoT(E319Q) mutant, expression from the reporter construct spanning P1 was somewhat lower in late logarithmic phase (OD600 = 1.0) but not at later times (OD600 ≥ 1.5) (Fig. (Fig.2C).2C). In contrast, the reporter construct spanning P2 showed no significant change when it was introduced into the relA or spoT mutant backgrounds (Fig. (Fig.2D2D).

Both iraD transcripts are positively influenced by (p)ppGpp, in a growth phase dependent manner.

We examined the (p)ppGpp regulation of iraD expression by Northern blot analysis. We were able to detect two transcripts, corresponding in length to those derived from promoter regions P1 and P2 through the iraD coding region. The longer transcript, presumably from distal promoter P1, was only detected in stationary phase cells. In agreement with the luciferase reporter assays, Northern blot analysis of the iraD transcripts showed positive regulation by (p)ppGpp. However, in contrast to the reporter assays, the Northern blot data revealed potential regulation by (p)ppGpp on the proximal promoter P2, in addition to the distal promoter P1. In exponential phase, levels of the 500-nt transcript were reduced dramatically by a mutation in relA, but less so by the spoT(E319Q) allele (Fig. (Fig.3);3); levels in the relA spoT double mutant were somewhat lower than relA alone. However, in stationary phase, the 500-nt transcript was not diminished, even in the relA spoT background. The larger stationary phase-specific transcript was virtually absent in the relA spoT double mutant (Fig. (Fig.3),3), although not significantly diminished by either relA or spoT(E319Q) alone.

FIG. 3.
Northern blot analysis of iraD transcripts in the MG1655, spoT(E319Q), relAΔ::FRT, and relAΔ spoTΔ strains. Samples were prepared from cells grown to either an OD of 0.7 (±0.1) or 1.6 (±0.1) as indicated. The graphs ...

RNA levels in Northern blot analysis confirmed the regulation of iraD transcripts by (p)ppGpp as apparent from the luciferase reporter assays, but indicated some discrepancies between the two that indicate a potentially more complex pattern of regulation. In particular, high levels of the smaller transcript did not correspond with high expression with the luciferase fusion. Whereas both small and large transcripts were almost equally abundant in stationary phase, expression from the fusion constructs appeared to be driven exclusively by the distal P1 promoter during stationary phase. The discrepancy between the two techniques may be due to postinitiation regulation of the transcripts, reflected in the reporter data but not at the level of iraD mRNA; alternatively, the reporter operon may be more sensitive to initiation or elongation effects specific to P2. Interestingly, each of the two transcripts had differential regulation with growth phase and responded to regulators of (p)ppGpp levels, although in different ways.

DksA is a negative regulator of iraD transcription.

DksA is a small protein that binds RNA polymerase and alters open complex formation on (p)ppGpp-regulated genes (34, 35). In vivo, DksA is required for the induction or repression of genes during the stringent response to amino acid starvation. We investigated a possible role for DksA in regulation of iraD by first examining expression of the full-length fusion of the upstream region (iraD::luxCDABE) in a dksA-null mutant. Surprisingly, the expression of iraD::luxCDABE was enhanced between 10 to 100-fold, depending on the growth phase of the cells, suggesting that DksA acts as a repressor of iraD expression (Fig. (Fig.4A).4A). Overexpression of DksA from a high copy plasmid (data not shown), led to lower levels of iraD::luxCDABE expression, even in relA spoT strains, confirming a negative role of DksA on IraD expression.

FIG. 4.
Expression of the full-length IraD promoter fusion (iraD::luxCDABE) in strains lacking DksA. Each data point is an average of four independent determinations. The variability is shown with error bars in both graphs. RLU, relative luminescence units (bioluminescence ...

Examination of iraD mRNA by Northern blots confirmed a repressive effect of DksA. In exponential phase, the 500-nt transcript was in greater abundance in the dksA mutant strain (Fig. (Fig.4D).4D). In stationary phase, the level of the larger 800-nt transcript was likewise higher in the dksA mutant.

A stringent RNA polymerase highly induces iraD expression.

To confirm the effect of the stringent response on iraD, we used a mutated version of RNA polymerase that mimics the “stringent” state (49). Interactions between stringent promoters and RNA polymerase are intrinsically unstable, and certain point mutations in the beta subunit of RNA polymerase (RpoB) core transcription complex exacerbate these interactions, thereby mimicking RNA polymerase bound to ppGpp during the stringent response (49). These “stringent” rpoB alleles reduce initiation at negatively regulated stringent promoters, such as those for rRNA, and increase expression of positively regulated stringent promoters. Therefore, if iraD is regulated by ppGpp, we would expect to see higher levels of expression from the iraD::lux reporter in any one of these stringent rpoB backgrounds.

We obtained one such allele, rpoB3370(T563P) (referred to below as rpoB*), and looked at iraD's expression in that strain background. As expected, iraD::lux CDABE was highly induced, by as much as 4 orders of magnitude, throughout the early, mid, and late exponential growth phases in the rpoB* background relative to wild-type strains, supporting the model that iraD is stringently regulated (Fig. (Fig.5A).5A). Similar induction by rpoB* was seen for fusions carrying either P1 or P2 promoter regions (Fig. 5D and E), indicating that both promoters are probably regulated by the stringent response.

FIG. 5.
Effects of a stringent rpoB allele on iraD transcription and RpoS levels. (A) iraD::luxCDABE expression from a −600 to −1 fusion construct in a “stringent” RNA polymerase mutant. Each data point is an average of three independent ...

Induction of IraD by a stringent RNAP has consequences on RpoS levels.

IraD appears a determinant of RpoS levels and stability (29) during normal exponential growth and, as shown here, during the transition to the stationary phase. To determine how much of the stringent response effect on RpoS was mediated through iraD, we examined RpoS steady-state levels and stability by Western blot analysis. Stability could be specifically ascertained by the addition of chloramphenicol to block new synthesis, followed by Western blotting with RpoS antibody. The effect of the rpoB* mutation should increase and/or stabilize RpoS levels through IraD, IraP, and perhaps other antiadaptors (9, 11, 29).

As expected, there were higher levels of RpoS in the rpoB* strain compared to the control wild-type strain (Fig. 5B and C). The elevated levels of RpoS protein detected in this strain were iraD dependent; this was apparent even when chloramphenicol was added to prevent new RpoS synthesis, supporting the fact that IraD effects on RpoS are posttranslational. A residual amount of RpoS was stabilized in rpoB* strains compared to wild-type, even after the deletion of iraD, probably through the action of other stringently regulated antiadaptors such as IraP (10).

DNA damage induction of iraD is not dependent on the stringent response and is mediated by the distal promoter.

We showed previously that iraD is highly induced after DNA damage by AZT or oxidative stress and that this induction is independent of the SOS response (29). To determine a role for (p)ppGpp in this induction, we assayed the iraD::luxCDABE reporter in the absence of (p)ppGpp or dksA. After AZT treatment, iraD::luxCDABE was induced to wild-type levels in (p)ppGpp0 strains, implying that (p)ppGpp is not necessary for the induction of iraD after DNA damage (Fig. (Fig.6A).6A). In the dksA-null strain, the induction after AZT treatment was higher than that of the wild-type, probably due to the derepression of the promoter in these backgrounds as we observed earlier. We did not observe any induction in the rpoB* strain background, probably because the gene is maximally induced even before AZT treatment (Fig. (Fig.5A5A).

FIG. 6.
Expression of various iraD::luxCDABE reporters in response to DNA damage. (A) Gray bars show expression in response to H2O (Ctrl), and black bars show expression in response to 1 μg of AZT/ml for at least two isolates. RLU, relative luminescence ...

Since we observed differences in the growth-phase regulation of the two transcripts of iraD, we wondered whether the DNA damage induction was also different between the two promoters. To address this using the luciferase fusion constructs, we looked at induction by AZT and peroxide of either the distal or the proximal promoter regions compared to the wild type. Interestingly, the distal promoter fusion containing the P1 start site showed strong induction with AZT treatment as well as with peroxide treatment, whereas the proximal promoter fusion containing the P2 start site did not (Fig. (Fig.6B).6B). This suggests that, similar to growth-phase regulation, the regulation of iraD expression in response to DNA damage may be predominantly at the distal promoter. The biological readout of iraD expression, accumulation of RpoS after DNA damage and during the transition to stationary phase, therefore appears to correlate with expression from the distal promoter. The role of the proximal promoter, if any, to iraD expression remains to be determined.


Expression of iraD is (p)ppGpp dependent.

Expression of iraD, as detected by a luciferase fusion, varies dramatically during the growth of cultures and is induced as cells begin to starve and enter the stationary phase of growth in rich medium. Similar to another antiadaptor, IraP, IraD expression seems to be regulated by (p)ppGpp, the signaling molecule of the stringent response, produced by either RelA and SpoT. In RNA polymerase mutants that mimic the stringent transcriptional state even in the absence of ppGpp production (49), iraD expression as detected by luciferase fusions was dramatically elevated, particularly in the exponential phase of growth when its levels increased by more than 4 orders of magnitude. In this rpoB* mutant, growth phase effects on iraD expression were minimized, with expression high throughout. This suggests that (p)ppGpp has the potential to modulate IraD not only during the transition to stationary phase, when its levels begin to accumulate, but also very sensitively in the mid-exponential phase of growth, when (p)ppGpp production is presumably low or transient in nature. This regulation by (p)ppGpp may be either direct or indirect via other regulatory factors, although we note that iraD putative promoter P1 has a AT-rich “discriminator” region downstream of the −10 sequence, a finding consistent with direct positive regulation by ppGpp.

Levels of ppGpp (which predominates over pppGpp) increase in response to a number of cellular stresses, the best studied of which is amino acid starvation, and naturally increase during late stages of exponential growth when nutrients begin to be depleted (reviewed in references 13 and 37). RelA, a ribosome-associated factor, is responsible for (p)ppGpp synthesis during translational stalling that accompanies amino acid starvation; SpoT has been associated with ppGpp synthesis associated with energy stress, such as carbon starvation.

As detected by promoter fusions to the luciferase operon, iraD expression was dependent on (p)ppGpp production during all phases of growth, with a reduction of >10-fold at the mid-exponential phase of growth and ~5-fold in the stationary phase. Surprisingly, the low level of iraD expression seen during exponential phase was reduced by both the loss of RelA and the ppGpp synthetase activity of SpoT (specifically affected by the SpoT E319Q mutation), indicating that both enzymes are required for maximal IraD expression in this growth phase. This is a puzzling result, since SpoT and RelA act independently and respond to different stimuli; this result may be explained by upregulation of SpoT (p)ppGpp hydrolase activity in the spoT(E319Q) mutant, causing a decrease in RelA-dependent (p)ppGpp synthesis in this mutant. In contrast, high levels of expression of iraD as cells were in transition to stationary phase were dependent on (p)ppGpp production by either RelA or SpoT, with a strong reduction seen only in the double mutant strain. Nutrient depletion in rich medium, such as that used in the present study, is likely to involve a complex combination of factors, with contributions to (p)ppGpp production by both RelA and SpoT.

Although both IraP and IraD antiadaptors appear to be regulated by the stringent response, there is a notable difference. The effect of ppGpp on IraP appear to be restricted to an integrated signal with phosphate starvation (10). In contrast, the effects of the stringent response on iraD appeared to be independent of its induction through DNA damage.

Furthermore, our results show that IraD is a determinant of RpoS levels during the transition to stationary phase, when high levels of (p)ppGpp begin to accumulate. An allele of rpoB, encoding the beta subunit of RNA polymerase, that causes the enzyme to display stringent transcription characteristics (49) constitutively caused a dramatic increase in iraD expression and elevated levels of RpoS levels, partially dependent on IraD. In the absence of IraD, RpoB* mutants had lower levels of RpoS, which were nonetheless stable, suggesting that other antiadaptors, including IraP, may substitute for IraD but with limited capacity relative to IraD.

DksA inhibits IraD expression.

DksA, a small transcription factor involved in the stringent response, acted primarily as a negative regulator of iraD through mid-exponential growth, and these effects were diminished in stationary phase. This effect could be direct or indirect.

Surprisingly, we found that, unlike many stringently regulated genes, the regulatory effects of ppGpp and DksA on iraD are opposing: (p)ppGpp is a positive regulator, whereas DksA is a negative regulator of the gene. The opposing and independent relationship between DksA and (p)ppGpp in iraD's regulation is similar to that found in a previous study wherein (p)ppGpp promoted and DksA antagonized cell adhesion, presumably a consequence of effects on fimB expression (2), as cultures entered stationary phase (28). DksA and (p)ppGpp also have opposing effects on a number of genes, primarily involved in chemotaxis and motility (1). The iraD gene was not found among this set but may have been missed because of growth condition differences or due to difficulty in the detection of the iraD transcript. The mechanism of this antagonistic and (p)ppGpp-independent effect by DksA on some genes appears to be competition with the GreA/B elongation factors (1, 2).

IraD transcripts.

Northern blot analysis confirmed features revealed by the gene fusion studies and presented new complexities. We detected two primary transcripts in wild-type strains: an ~500-nt RNA in exponential phase and an 800-nt transcript present in the stationary phase of growth. We mapped the start sites corresponding to these transcripts by sequence analysis and 5′ RACE at positions −137 and −417 relative to the start codon of iraD, which are juxtaposed to identifiable promoter sequences, P2 and P1, respectively. Analyses of these regions fused to luciferase are consistent with these as the two promoters for iraD, with the distal promoter induced strongly during stationary phase.

There are some discrepancies between the full-length fusion reporter and RNA analysis. One difference is that the luciferase fusion exhibited an increase of many orders of magnitude in expression in late exponential phase, whereas the Northern analysis indicates a modest increase in the total RNA levels, with a dramatic increase only in the larger transcript. Reporter fusions suggested that only the distal promoter, P1, contributed to late-exponential-phase induction of iraD. This suggests that the larger 800-nt transcript emanating from P1 has properties that improve the elongation, stability, or translatability of the mRNA as reflected in luciferase production. This property of the larger transcript is also apparent in the partial fusion reporter containing the distal promoter P1 compared to the proximal promoter P2.

A need for posttranscriptional activation of the P2 transcript could explain a number of the observed discrepancies between the reporter data and Northern blots. This includes reduction of transcript levels by RelA, but with no reduction in expression detected by the P2 luciferase reporter construct. An additional mode of regulation of iraD in the exponential phase could promote a rapid, sensitive, and potentially transient activation during this growth phase.

Our results suggest that the growth-phase regulation of iraD is mediated by a combination of (p)ppGpp-dependent positive regulation and DksA-dependent negative regulation and possibly other factors. Even in (p)ppGpp0 strains, iraD expression, as detected by the reporter fusions, was growth phase dependent, indicating potentially an additional mode of growth regulation. Both P1 and P2 transcripts were positively regulated by (p)ppGpp, although the small transcript became (p)ppGpp independent during the stationary phase. The small transcript required RelA predominantly during exponential growth, whereas the larger requires either RelA or SpoT during the transition to stationary phase. This suggests that (p)ppGpp production is primarily RelA dependent in growing cells but switches to more balanced production by both RelA and SpoT in the late exponential phase. Both transcripts were repressed by DksA, with higher levels in dksA mutants. Again, as with (p)ppGpp, the smaller transcript became resistant to DksA effects during the stationary phase, indicating a switch in its regulation.

Connection between stringent response and RpoS through IraD.

We observed IraD-dependent stabilization of RpoS during normal growth and in stringent-mimic RpoB* cells. The accumulation of RpoS as cells enter the stationary phase was primarily dependent on IraD. A previous report (12) did not detect stabilization of RpoS during accumulation of (p)ppGpp, although we note that this latter experiment was performed with different growth conditions (minimal versus rich medium growth), and the reported RpoS half-life was much longer than what we observed (5 min versus <1 min). RpoS accumulation with (p)ppGpp may involve a number of influences, including both increases in translational efficiency, as suggested in the previous study, and protein stability, as suggested by our study. The only known biochemical function of IraD is to bind the adaptor RssB that targets RpoS for proteolysis by ClpXP (9); however, it is also conceivable that IraD either directly or indirectly affects other factors that influence RpoS levels.

Both RpoS and (p)ppGpp lead to major changes in gene expression in response to cellular stress and different growth conditions. These two systems may be connected through the regulation of RpoS at many levels, including transcriptional, translational, and protein stability. Our results raise the possibility that (p)ppGpp may regulate RpoS, at least in part, through effects on expression of the stabilizing factor, IraD.


We thank Michael Cashel, Ding Jin, Hirotada Mori, and Tina van Dyk for strains and plasmids.

This study was supported by grants from the General Medical Sciences Institute (GM051753 to S.T.L. and T32 GM007122 to A.E.F.).


[down-pointing small open triangle]Published ahead of print on 9 October 2009.


1. Aberg, A., J. Fernández-Vázquez, J. D. Cabrer-Panes, A. Sánchez, and C. Balsalobre. 2009. Similar and divergent effects of ppGpp and DksA deficiencies on transcription in Escherichia coli. J. Bacteriol. 191:3226-3236. [PMC free article] [PubMed]
2. Aberg, A., V. Shingler, and C. Balsalobre. 2008. Regulation of the fimB promoter: a case of differential regulation by ppGpp and DksA in vivo. Mol. Microbiol. 67:1223-1241. [PubMed]
3. Barker, M. M., T. Gaal, C. A. Josaitis, and R. L. Gourse. 2001. Mechanism of regulation of transcription initiation by ppGpp. I. Effects of ppGpp on transcription initiation in vivo and in vitro. J. Mol. Biol. 305:673-688. [PubMed]
4. Barker, M. M., and R. L. Gourse. 2001. Regulation of rRNA transcription correlates with nucleoside triphosphate sensing. J. Bacteriol. 183:6315-6323. [PMC free article] [PubMed]
5. Bearson, S. M., W. H. Benjamin, Jr., W. E. Swords, and J. W. Foster. 1996. Acid shock induction of RpoS is mediated by the mouse virulence gene mviA of Salmonella typhimurium. J. Bacteriol. 178:2572-2579. [PMC free article] [PubMed]
6. Blattner, F. R., G. Plunkett III, C. A. Bloch, N. T. Perna, V. Burland, M. Riley, J. Collado-Vides, J. D. Glasner, C. K. Rode, G. F. Mayhew, J. Gregor, N. W. Davis, H. A. Kirkpatrick, M. A. Goeden, D. J. Rose, B. Mau, and Y. Shao. 1997. The complete genome sequence of Escherichia coli K-12. Science 277:1453-1474. [PubMed]
7. Block, R., and A. W. Haseltine. 1975. Purification and properties of stringent factor. J. Biol. Chem. 250:1212-1217. [PubMed]
8. Bouche, S., E. Klauck, D. Fischer, M. Lucassen, K. Jung, and R. Hengge-Aronis. 1998. Regulation of RssB-dependent proteolysis in Escherichia coli: a role for acetyl phosphate in a response regulator-controlled process. Mol. Microbiol. 27:787-795. [PubMed]
9. Bougdour, A., C. Cunning, P. J. Baptiste, T. Elliott, and S. Gottesman. 2008. Multiple pathways for regulation of σS (RpoS) stability in Escherichia coli via the action of multiple anti-adaptors. Mol. Microbiol. 68:298-313. [PubMed]
10. Bougdour, A., and S. Gottesman. 2007. ppGpp regulation of RpoS degradation via anti-adaptor protein IraP. Proc. Natl. Acad. Sci. USA 104:12896-12901. [PubMed]
11. Bougdour, A., S. Wickner, and S. Gottesman. 2006. Modulating RssB activity: IraP, a novel regulator of σS stability in Escherichia coli. Genes Dev. 20:884-897. [PubMed]
12. Brown, L., D. Gentry, T. Elliott, and M. Cashel. 2002. DksA affects ppGpp induction of RpoS at a translational level. J. Bacteriol. 184:4455-4465. [PMC free article] [PubMed]
13. Cashel, M., D. R. Gentry, V. H. Hernandez, and D. Vinella. 1996. The stringent response, p. 1458-1496. In F. C. Neidhardt, R. Curtiss III, J. L. Ingraham, E. C. C. Lin, K. B. Low, B. Magasanik, W. S. Reznikoff, M. Riley, M. Schaechter, and H. E. Umbarger (ed.), Escherichia coli and Salmonella: cellular and molecular biology, 2nd ed. ASM Press, Washington, DC.
14. Datsenko, K. A., and B. L. Wanner. 2000. One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc. Natl. Acad. Sci. USA 97:6640-6645. [PubMed]
15. Dong, T., M. G. Kirchhof, and H. E. Schellhorn. 2008. RpoS regulation of gene expression during exponential growth of Escherichia coli K12. Mol. Genet. Genomics 279:267-277. [PubMed]
16. Dong, T., and H. E. Schellhorn. 2009. Control of RpoS in global gene expression of Escherichia coli in minimal media. Mol. Genet. Genomics 281:19-33. [PubMed]
17. Gentry, D. R., V. J. Hernandez, L. H. Nguyen, D. B. Jensen, and M. Cashel. 1993. Synthesis of the stationary-phase sigma factor sigma s is positively regulated by ppGpp. J. Bacteriol. 175:7982-7989. [PMC free article] [PubMed]
18. Harinarayanan, R., H. Murphy, and M. Cashel. 2008. Synthetic growth phenotypes of Escherichia coli lacking ppGpp and transketolase A (tktA) are due to ppGpp-mediated transcriptional regulation of tktB. Mol. Microbiol. 69:882-894. [PMC free article] [PubMed]
19. Hengge-Aronis, R. 1996. Back to log phase: sigma S as a global regulator in the osmotic control of gene expression in Escherichia coli. Mol. Microbiol. 21:887-893. [PubMed]
20. Hengge-Aronis, R. 2002. Recent insights into the general stress response regulatory network in Escherichia coli. J. Mol. Microbiol. Biotechnol. 4:341-346. [PubMed]
21. Hengge-Aronis, R. 2002. Signal transduction and regulatory mechanisms involved in control of the σS (RpoS) subunit of RNA polymerase. Microbiol. Mol. Biol. Rev. 66:373-395. [PMC free article] [PubMed]
22. Hernandez, V. J., and H. Bremer. 1991. Escherichia coli ppGpp synthetase II activity requires spoT. J. Biol. Chem. 266:5991-5999. [PubMed]
23. Klauck, E., A. Typas, and R. Hengge. 2007. The sigmaS subunit of RNA polymerase as a signal integrator and network master regulator in the general stress response in Escherichia coli. Sci. Prog. 90:103-127. [PubMed]
24. Lacour, S., and P. Landini. 2004. SigmaS-dependent gene expression at the onset of stationary phase in Escherichia coli: function of σS-dependent genes and identification of their promoter sequences. J. Bacteriol. 186:7186-7195. [PMC free article] [PubMed]
25. Lange, R., D. Fischer, and R. Hengge-Aronis. 1995. Identification of transcriptional start sites and the role of ppGpp in the expression of rpoS, the structural gene for the sigma S subunit of RNA polymerase in Escherichia coli. J. Bacteriol. 177:4676-4680. [PMC free article] [PubMed]
26. Lange, R., and R. Hengge-Aronis. 1994. The cellular concentration of the sigma S subunit of RNA polymerase in Escherichia coli is controlled at the levels of transcription, translation, and protein stability. Genes Dev. 8:1600-1612. [PubMed]
27. Loewen, P. C., and R. Hengge-Aronis. 1994. The role of the sigma factor sigma S (KatF) in bacterial global regulation. Annu. Rev. Microbiol. 48:53-80. [PubMed]
28. Magnusson, L. U., B. Gummesson, P. Joksimovic, A. Farewell, and T. Nystrom. 2007. Identical, independent, and opposing roles of ppGpp and DksA in Escherichia coli. J. Bacteriol. 189:5193-5202. [PMC free article] [PubMed]
29. Merrikh, H., A. E. Ferrazzoli, A. Bougdour, A. Olivier-Mason, and S. T. Lovett. 2009. A DNA damage response in Escherichia coli involving the alternative sigma factor, RpoS. Proc. Natl. Acad. Sci. USA 106:611-616. [PubMed]
30. Moreno, M., J. P. Audia, S. M. Bearson, C. Webb, and J. W. Foster. 2000. Regulation of sigma S degradation in Salmonella enterica var. typhimurium: in vivo interactions between sigma S, the response regulator MviA(RssB) and ClpX. J. Mol. Microbiol. Biotechnol. 2:245-254. [PubMed]
31. Muffler, A., D. Fischer, S. Altuvia, G. Storz, and R. Hengge-Aronis. 1996. The response regulator RssB controls stability of the σS subunit of RNA polymerase in Escherichia coli. EMBO J. 15:1333-1339. [PubMed]
32. Murray, K. D., and H. Bremer. 1996. Control of spoT-dependent ppGpp synthesis and degradation in Escherichia coli. J. Mol. Biol. 259:41-57. [PubMed]
33. Patten, C. L., M. G. Kirchhof, M. R. Schertzberg, R. A. Morton, and H. E. Schellhorn. 2004. Microarray analysis of RpoS-mediated gene expression in Escherichia coli K-12. Mol. Genet. Genomics 272:580-591. [PubMed]
34. Paul, B. J., M. M. Barker, W. Ross, D. A. Schneider, C. Webb, J. W. Foster, and R. L. Gourse. 2004. DksA: a critical component of the transcription initiation machinery that potentiates the regulation of rRNA promoters by ppGpp and the initiating NTP. Cell 118:311-322. [PubMed]
35. Paul, B. J., M. B. Berkmen, and R. L. Gourse. 2005. DksA potentiates direct activation of amino acid promoters by ppGpp. Proc. Natl. Acad. Sci. USA 102:7823-7828. [PubMed]
36. Peterson, C. N., N. Ruiz, and T. J. Silhavy. 2004. RpoS proteolysis is regulated by a mechanism that does not require the SprE (RssB) response regulator phosphorylation site. J. Bacteriol. 186:7403-7410. [PMC free article] [PubMed]
37. Potrykus, K., and M. Cashel. 2008. (p)ppGpp: still magical? Annu. Rev. Microbiol. 62:35-51. [PubMed]
38. Pratt, L. A., and T. J. Silhavy. 1996. The response regulator SprE controls the stability of RpoS. Proc. Natl. Acad. Sci. USA 93:2488-2492. [PubMed]
39. Pruteanu, M., and R. Hengge-Aronis. 2002. The cellular level of the recognition factor RssB is rate-limiting for σS proteolysis: implications for RssB regulation and signal transduction in σS turnover in Escherichia coli. Mol. Microbiol. 45:1701-1713. [PubMed]
40. Schweder, T., K. H. Lee, O. Lomovskaya, and A. Matin. 1996. Regulation of Escherichia coli starvation sigma factor (σS) by ClpXP protease. J. Bacteriol. 178:470-476. [PMC free article] [PubMed]
41. Tanaka, K., Y. Takayanagi, N. Fujita, A. Ishihama, and H. Takahashi. 1993. Heterogeneity of the principal sigma factor in Escherichia coli: the rpoS gene product, sigma 38, is a second principal sigma factor of RNA polymerase in stationary-phase Escherichia coli. Proc. Natl. Acad. Sci. USA 90:3511-3515. [PubMed]
42. Teich, A., S. Meyer, H. Y. Lin, L. Andersson, S. Enfors, and P. Neubauer. 1999. Growth rate related concentration changes of the starvation response regulators σS and ppGpp in glucose-limited fed-batch and continuous cultures of Escherichia coli. Biotechnol. Prog. 15:123-129. [PubMed]
43. Troutt, A. B., M. G. McHeyzer-Williams, B. Pulendran, and G. J. Nossal. 1992. Ligation-anchored PCR: a simple amplification technique with single-sided specificity. Proc. Natl. Acad. Sci. USA 89:9823-9825. [PubMed]
44. Van Dyk, T. K., Y. Wei, M. K. Hanafey, M. Dolan, M. J. Reeve, J. A. Rafalski, L. B. Rothman-Denes, and R. A. LaRossa. 2001. A genomic approach to gene fusion technology. Proc. Natl. Acad. Sci. USA 98:2555-2560. [PubMed]
45. Weber, H., T. Polen, J. Heuveling, V. F. Wendisch, and R. Hengge. 2005. Genome-wide analysis of the general stress response network in Escherichia coli: σS-dependent genes, promoters, and sigma factor selectivity. J. Bacteriol. 187:1591-1603. [PMC free article] [PubMed]
46. Xiao, H., M. Kalman, K. Ikehara, S. Zemel, G. Glaser, and M. Cashel. 1991. Residual guanosine 3′,5′-bispyrophosphate synthetic activity of relA-null mutants can be eliminated by spoT-null mutations. J. Biol. Chem. 266:5980-5990. [PubMed]
47. Zhou, Y., and S. Gottesman. 1998. Regulation of proteolysis of the stationary-phase sigma factor RpoS. J. Bacteriol. 180:1154-1158. [PMC free article] [PubMed]
48. Zhou, Y., S. Gottesman, J. R. Hoskins, M. R. Maurizi, and S. Wickner. 2001. The RssB response regulator directly targets σS for degradation by ClpXP. Genes Dev. 15:627-637. [PubMed]
49. Zhou, Y. N., and D. J. Jin. 1998. The rpoB mutants destabilizing initiation complexes at stringently controlled promoters behave like “stringent” RNA polymerases in Escherichia coli. Proc. Natl. Acad. Sci. USA 95:2908-2913. [PubMed]

Articles from Journal of Bacteriology are provided here courtesy of American Society for Microbiology (ASM)