|Home | About | Journals | Submit | Contact Us | Français|
The Escherichia coli guaB promoter (PguaB) is responsible for directing transcription of the guaB and guaA genes, which specify the biosynthesis of the nucleotide GMP. PguaB is subject to growth rate-dependent control (GRDC) and possesses an UP element that is required for this regulation. In addition, PguaB contains a discriminator, three binding sites for the nucleoid-associated protein FIS, and putative binding sites for the regulatory proteins DnaA, PurR, and cyclic AMP receptor protein (CRP). Here we show that the CRP-cyclic AMP (cAMP) complex binds to a site located over 100 bp upstream of the guaB transcription start site, where it serves to downregulate PguaB. The CRP-mediated repression of PguaB activity increases in media that support lower growth rates. Inactivation of the crp or cyaA gene or ablation/translocation of the CRP site relieves repression by CRP and results in a loss of GRDC of PguaB. Thus, GRDC of PguaB involves a progressive increase in CRP-mediated repression of the promoter as the growth rate decreases. Our results also suggest that the CRP-cAMP complex does not direct GRDC at PguaB and that at least one other regulatory factor is required for conferring GRDC on this promoter. However, PurR and DnaA are not required for this regulatory mechanism.
The Escherichia coli guaB promoter (PguaB) regulates transcription of the guaBA operon. The guaB and guaA genes encode IMP dehydrogenase and GMP synthetase, respectively, and are required for synthesis de novo of GMP from the common purine precursor IMP (46, 75). PguaB responds to a variety of physiological signals. For example, the activity of PguaB increases as a function of the cellular growth rate, such that guaBA mRNA forms an increasing fraction of total cell mass at higher growth rates (16, 33, 34). This form of regulation is referred to as growth rate-dependent control (GRDC) (17, 27). PguaB is also subject to stringent control (16, 71), growth phase-dependent regulation (34), and purine repression (16, 70), and its activity is coupled to the DNA replication cycle (73).
The multivalent regulation of PguaB is reflected by the presence of a number of cis-acting regulatory sites that overlap this promoter. An UP element, located immediately upstream of the promoter −35 region, strongly enhances transcription and is required for GRDC of this promoter (28, 33). Three binding sites for the nucleoid-associated protein FIS have been identified, centered near positions −11, +8, and +29 relative to the guaB transcription start site. Accordingly, FIS has been shown to repress transcription from PguaB in vitro. However, FIS is not required for GRDC of PguaB (34). PguaB also contains a putative binding site for PurR that overlaps the core promoter region (16, 31) (see Fig. Fig.1).1). Consistent with this, IMP dehydrogenase activity is higher in a purR mutant strain, and unlike the situation in the wild-type strain, this activity is not repressed by growth in the presence of high concentrations of guanine or guanosine (46, 48). Approximately 200 bp downstream from the translation initiation codon for guaB is a consensus DnaA binding site that has been shown to bind DnaA in vitro and is required for DnaA-mediated repression of guaB transcription in vivo (73, 74). A second, nonconsensus DnaA site overlaps the guaB promoter. Although it does not bind DnaA in vitro, it may be required in concert with the downstream site for efficient DnaA-mediated downregulation of PguaB in vivo (73, 74).
The cyclic AMP receptor protein (CRP) is a cyclic AMP (cAMP)-dependent global transcription regulator (for reviews, see references 11 and 45). It has long been established that in the absence of exogenous glucose, CRP activates the expression of a large number of genes required for catabolism of alternative carbon sources by E. coli (for a review, see reference 41). More recent genomic studies have shown that CRP also activates transcription of genes that encode enzymes involved in central carbon metabolism and transporters of various alternative carbon sources (26, 29, 85). CRP activates transcription by binding to specific sites located upstream of promoters and contacting RNAP (11). Promoters that utilize CRP as the sole activator are categorized as either class I or class II promoters (11). At class I CRP-dependent promoters, homodimeric CRP binds DNA sites centered near positions −61.5, −71.5, −82.5, and −92.5 with respect to the transcription start site, and the downstream CRP monomer stimulates transcription by contacting the C-terminal domain of the RNAP α subunit (αCTD) through a surface-exposed loop known as activating region 1 (AR1; residues 156 to 164). At class II CRP-dependent promoters, CRP binds to a site centered near position −41.5, and transcription is stimulated through interactions between αCTD and AR1 of the upstream CRP monomer and between αNTD and an additional surface, referred to as activating region 2 (AR2; residues 19, 21, 96, and 101), on the downstream CRP monomer (11, 24, 76). At class III promoters, optimum transcription activation is achieved by the binding of at least two CRP dimers or a combination of CRP and another regulatory protein(s) (11). CRP has also been shown to repress transcription of some genes, including the crp gene itself and cyaA, encoding adenylate cyclase (2, 25, 26, 52, 85). Transcription repression by CRP can occur through the occupation of DNA sites that overlap core promoter regions (69) or by stabilization of a transcriptional repressor or corepressor bound to the promoter region (53, 77).
A putative binding site for CRP has been identified centered at position −117.5 relative to the start site for guaB transcription (Fig. (Fig.1).1). The CRP site matches the consensus at 17/22 positions, including 9/10 positions in the core binding motifs that are critical for CRP binding (9, 30, 35). Consistent with this, CRP has been shown to bind to an ~300-bp DNA fragment that extends to position −253 with respect to the PguaB transcription start site (35). It was proposed that CRP functions as an activator at PguaB. In a separate transcriptomic study, guaB was identified as one of a number of genes that are subject to “CRP-dependent glucose activation” (26). We have previously shown that sequences located between positions −133 and −100, which include the putative CRP site, are required for GRDC of PguaB (33). To resolve the apparent inconsistencies between some of these observations, we carried out an investigation into the role of CRP at PguaB. Here we show that CRP binds to the putative CRP site located upstream of PguaB, whereupon it serves to decrease promoter activity. Moreover, occupancy of this site by CRP is required for GRDC of PguaB.
Bacterial strains and plasmids used in this study are listed in Table Table1.1. Promoter fragments were constructed by standard PCR techniques. Oligonucleotide primer sequences are shown in Table S1 in the supplemental material. For promoter activity measurements in vivo, strains containing single-copy promoter-lacZ transcriptional fusions were employed. All transcriptional fusions were carried on λ prophages and were constructed in the VH1000 genetic background, using a system based on λimm21 (59, 68). Apart from PguaB (−253 to +10), all PguaB-lacZ fusions contained downstream endpoints at position +36 with respect to the guaB transcription start site. The cya1400::kan, crp::cat, and purR6::Tn10 alleles were transferred into lysogenic VH1000 derivative strains by P1 transduction (51).
Logarithmically growing cells containing chromosomally integrated PguaB-lacZ transcriptional fusions were employed for the measurement of promoter activity as a function of growth rate. Cells from overnight cultures grown in medium supporting the lowest growth rate were inoculated to a starting optical density at 600 nm (OD600) of 0.02 into different media that supported a range of growth rates and were grown with aeration at 37°C. The set of culture media used for growing strains containing wild-type crp and cyaA alleles was based on M9 minimal medium and is referred to here as standard media (33). The set of media used to grow strains containing deletions in the crp and/or cyaA gene is referred to as CRP media and was based on M9 minimal medium containing one of the following (in order of increasing growth rates supported for wild-type strains): 0.4% (wt/vol) fructose, 0.4% (wt/vol) glucose, 0.4% (wt/vol) fructose plus 20 amino acids, 0.4% (wt/vol) fructose plus 1% (wt/vol) Casamino Acids, 0.4% (wt/vol) glucose plus 20 amino acids, or 0.4% (wt/vol) glucose plus 0.8% (wt/vol) Casamino Acids. The 20 amino acids comprised each amino acid at a final concentration of 20 μg/ml. All growth media also contained 5 μg/ml thiamine. Strains containing plasmids derived from pLG339 (i.e., pLG339ΔBS, pLG339CRP, pLG339CRP159L, and pLG339CRP101E) were grown in the presence of 25 μg/ml kanamycin, and strains containing plasmids derived from pBR322 (i.e., pDU9, pDCRP, and pHA7) were grown in the presence of 100 μg/ml ampicillin. cAMP sodium monohydrate (Sigma-Aldrich) was added to a final concentration of 5 mM, where included. Cultures were grown to an OD600 of 0.40 to 0.45, whereupon the β-galactosidase activity was measured following disruption of cells by sonication (51, 80). (To measure the effect of the ΔpurR allele on transcription from PguaB derivatives in bacteria growing in M9 medium containing glucose, cells were permeabilized by chloroform-sodium dodecyl sulfate [SDS] treatment.) All data points on GRDC plots represent the mean β-galactosidase activity (in Miller units) and mean growth rate for three independent experiments.
For measurement of transcription in vitro, supercoiled plasmid DNA was used, containing either the PguaB (−133 to +36) promoter, carried by plasmid pRLG770, or the synthetic class II CRP-dependent promoter CC(−41.5), carried by plasmid pSR. Multiple-round transcription reactions were performed as previously described, in the presence or absence of CRP, except that KCl was used at a concentration of 100 mM (33). Template was preincubated with 10 nM RNAP holoenzyme (Epicentre), 20 nM CRP, 200 μM cAMP, a 200 μM concentration (each) of CTP and either ATP or GTP, and 10 μM UTP for 10 min at 30°C. The reaction commenced after addition of the initiating nucleotide [200 μM GTP for PguaB and 200 μM ATP for the CC(−41.5) promoter] and was allowed to proceed for 10 min at 30°C. Reactions were terminated with an equal volume of stop solution (95% deionized formamide, 20 mM EDTA, 0.05% bromophenol blue, 0.05% xylene cyanol). Samples were fractionated in a 5.5% acrylamide gel containing 7 M urea, and transcript abundance was quantified using a FujiFilm FLA3000 phosphorimager.
DNA fragments containing PguaB sequences were amplified from pUC19 by a PCR using flanking pUC19-specific primers, as described previously (34). PCR products were digested with HindIII and purified by crush-soaking in a solution containing 0.2 M NaCl, 20 mM Tris-HCl (pH 8.0), and 1 mM EDTA (pH 8.0) following electrophoresis in a 6% polyacrylamide gel (50). Fragments were labeled at the HindIII end by use of [α-32P]dATP (3,000 Ci [1.11 × 1014 Bq]/mmol; MP Biomedicals) and the DNA polymerase I Klenow fragment. Labeled DNA (final concentration, 0.4 nM) was incubated at room temperature for 30 min in 10 μl of a buffer containing 20 mM HEPES (pH 8.0), 5 mM MgCl2, 50 mM potassium glutamate, 1 mM dithiothreitol, 10% (vol/vol) glycerol, 20 μg ml−1 sonicated calf thymus DNA (GE Healthcare), and, where appropriate, 200 μM cAMP, in the absence or presence of 200 nM CRP. Samples were loaded onto a 6% polyacrylamide gel (37.5:1 acrylamide:bisacrylamide) containing 7.5% (vol/vol) glycerol and 200 μM cAMP while running at ~15 V/cm and were electrophoresed for ~1 h at 4°C. Radiolabeled DNA was visualized using a FujiFilm FLA3000 phosphorimager.
DNase I footprinting was performed as described previously, using an EcoRI-XhoI DNA fragment isolated from plasmid pBSG-253 (extending from positions −253 to +36 of the guaB promoter) and labeled at the XhoI end of the template strand with [γ-32P]ATP (>7,000 Ci/mmol; MP Biomedicals), using T4 polynucleotide kinase (34). To facilitate CRP binding to DNA, reaction samples, each containing a final concentration of 2.5% (wt/vol) glycerol, were incubated with binding buffer (20 mM HEPES [pH 8.0], 5 mM MgCl2, 50 mM potassium glutamate, 1 mM dithiothreitol, 20 μg ml−1 sonicated calf thymus DNA) at room temperature for 30 min in the presence or absence of CRP at 200 nM and/or cAMP at 200 μM, and this was followed by digestion with DNase I. Samples were purified by phenol-chloroform extraction and ethanol precipitation, and DNA fragments were separated in a 6% polyacrylamide-7 M urea sequencing gel. A Maxam-Gilbert G+A sequencing ladder was run alongside the fragments. Footprints were visualized using a FujiFilm FLA3000 phosphorimager.
SDS-polyacrylamide gel electrophoresis (SDS-PAGE) was carried out using a 12% resolving gel (29:1 acrylamide-bisacrylamide). Western blotting was carried out to measure CRP levels in strain VH1000G-133. A derivative strain (VH1000G-133 Δcrp) that contained a deletion in the crp gene served as a negative control. For Western blotting, cells were grown in media supporting different growth rates, and the OD600 was measured periodically. Logarithmically growing cells at an OD600 of ~0.35 to 0.45 were harvested by centrifugation and disrupted by sonication, and the total protein concentration in the soluble fraction was determined using an RC DC protein assay kit (Bio-Rad). A total of 2.4 μg of total protein was fractionated by SDS-PAGE and transferred to a polyvinylidene difluoride membrane by electroblotting. The membrane was blocked with StartingBlock phosphate-buffered saline buffer (Pierce), and detection was performed using rabbit anti-CRP antiserum (a gift from H. Aiba), a secondary antibody conjugated to horseradish peroxidase (Caltag), and SuperSignal West Pico chemiluminescent substrate (Pierce). Protein bands corresponding to CRP were detected by autoradiography, and their intensities were quantified densitometrically.
DNase I footprinting was employed to determine whether the putative CRP site located at position −117.5 is bound by purified CRP. The results show that CRP protects a region from positions −107 to −128 relative to the guaB transcription start site, consistent with the location of the predicted CRP site (Fig. (Fig.2).2). No other sequences between positions −253 and +36 were protected by CRP (result not shown).
A PguaB promoter fragment with an upstream end point at position −133 and a downstream end point at position +36 with respect to the transcription start site [PguaB (−133 to +36)] contains all of the DNA sequence elements required for full GRDC of PguaB (Fig. (Fig.1)1) (33). To determine whether the CRP site centered at position −117.5 is required for GRDC, cells containing single-copy PguaB-lacZ fusions in which the CRP site was present [PguaB (−133 to +36)] or absent [PguaB (−117 to +36)] were grown at different growth rates, and the β-galactosidase activity was measured.
In agreement with previous observations, transcription from PguaB (−133 to +36) increased in medium supporting higher growth rates (16) (Fig. (Fig.3A).3A). As previously noted, this corresponded to an ~1.8-fold increase in promoter activity for a doubling of the growth rate (33). In contrast, the activity of PguaB (−117 to +36) was maintained at similar high levels at all growth rates tested (Fig. (Fig.3B).3B). The reason for this was due to a higher activity of PguaB (−117 to +36) than that of PguaB (−133 to +36) in bacteria growing in medium supporting low growth rates. This observation suggests that residues between positions −133 and −118, which include the CRP binding site, contribute in some way to GRDC of PguaB.
To determine whether the CRP binding site is required for GRDC of PguaB, promoter derivatives containing single-base-pair mutations in the CRP site were analyzed. CRP exhibits a high degree of specificity for a G base at position 7 (corresponding to a C at position 16) of the 22-bp consensus CRP site (aaaTGTGAtctagaTCACAttt) (the core binding site is shown in upper case, and positions 7 and 16 are underlined). Replacement of this residue with a C results in a large reduction in CRP binding to the consensus CRP site (30). Therefore, we introduced a G-to-C point mutation at the equivalent position in the PguaB CRP site (i.e., position −122 relative to the guaB transcription start site), giving rise to PguaB (−133 to +36, G-122C). Our results showed that introduction of this substitution led to a complete loss of GRDC (Fig. (Fig.3C).3C). Similar to the situation with PguaB (−117 to +36), abolition of GRDC was due to an increase in PguaB (−133 to +36, G-122C) activity relative to that of PguaB (−133 to +36), which was most pronounced at low growth rates. CRP binding to the mutant site was assessed by EMSA. This showed that the substitution caused a significantly reduced CRP-PguaB interaction (Fig. (Fig.3E).3E). Our results strongly suggest that binding of CRP to the site located at position −117.5 is required for repression of PguaB at low growth rates and that this plays a role in GRDC of PguaB.
In a complementary experiment, an A-to-C point mutation was introduced at position 18 in the PguaB CRP site [PguaB (−133 to +36, A-111C)], generating a consensus core CRP binding site (Fig. (Fig.1).1). Stronger binding of CRP to PguaB (−133 to +36, A-111C) was confirmed by EMSA (Fig. (Fig.3E).3E). The stronger binding of CRP to the consensus CRP site resulted in a 30 to 50% decrease in PguaB activity, depending on the growth rate (Fig. (Fig.3D).3D). However, the change in activity of this promoter as a function of the growth rate was not significantly different from that with PguaB (−133 to +36). Thus, whereas for the wild-type guaB promoter there was an ~1.8-fold increase in promoter activity as the growth rate doubled, an ~2-fold increase in activity was observed for the A-111C derivative. This result is consistent with our proposal that binding of CRP to the PguaB upstream region serves to downregulate transcription.
To confirm that GRDC of PguaB requires CRP, the activity of PguaB (−133 to +36) in relation to the growth rate was measured in a Δcrp strain. In the absence of CRP, PguaB exhibited an increase in activity at low growth rates relative to that in wild-type cells, and there was no longer a positive correlation between the promoter activity and the growth rate (Fig. (Fig.4A).4A). Moreover, as observed previously, the presence of the Δcrp allele resulted in a slow growth phenotype in all media tested (for example, see references 15, 18, 56, and 83). Normal GRDC of PguaB was restored following introduction of a low-copy-number plasmid specifying CRP (Fig. (Fig.4B).4B). These results are consistent with the observation that a functional CRP site is required for positive GRDC of PguaB.
To ascertain the requirement for the RNAP contact sites on CRP, plasmids harboring mutant crp alleles that encode derivatives containing inhibitory amino acid substitutions in AR1 or AR2 were introduced into the Δcrp strain harboring PguaB (−133 to +36), and the growth rate dependence of the promoter was measured. Control experiments with the CRP AR1 and AR2 variants demonstrated that this system could be used to measure the effects of inactivation of AR1 and AR2 at class I and class II CRP-dependent promoters in vivo in the same strain background (results not shown) (12, 81). Our results show that inactivation of AR1 resulted in a decrease in PguaB activity at all growth rates (Fig. (Fig.4C),4C), whereas disruption of AR2 function gave rise to a more modest decrease in the activity of this promoter (Fig. (Fig.4D).4D). However, the presence of either mutant CRP did not result in a significant change in GRDC of PguaB in comparison to that with wild-type CRP. This is more obvious in plots of relative promoter activity versus growth rate (see Fig. S1 in the supplemental material).
To test whether the location of the CRP site is important for repression of PguaB and for GRDC, the CRP site in PguaB (−133 to +36) was moved to a location centered at position −106.5 (approximately one helical turn downstream of the original location) and −128.5 (approximately one helical turn upstream of the original location). In each case, the original CRP site was disrupted by point mutations in the core region that are unfavorable for CRP binding (Table (Table1),1), and the repositioned CRP site retained the 22-bp sequence of the original. The activities of these guaB promoter derivatives in bacteria growing at different rates were then measured in a wild-type strain background. The results showed that moving the PguaB CRP site one helical turn upstream or downstream of the original location led to a complete loss of GRDC (Fig. 4E and F; also see Fig. S1 in the supplemental material). This suggests that the distance between the CRP site and the PguaB core promoter elements is important for repression of PguaB (and for GRDC) and/or that an adjacent or overlapping binding site for another transcription factor, disrupted by repositioning of the CRP site, may also play an important role in regulation of PguaB.
If the cAMP-CRP complex acts as the primary sensor of growth rate for PguaB, then an inverse correlation between the intracellular concentrations of CRP and/or cAMP and the cellular growth rate would be expected. For example, it was shown previously that the intracellular concentration of CRP is lower in E. coli cells growing in the presence of glucose than in cells utilizing glycerol as a sole carbon source (37). To examine whether the intracellular concentration of CRP varies with the growth rate, Western blotting was employed. The results showed that at growth rates of <1.0 doubling per hour, the intracellular concentration of CRP was approximately twofold higher than that at higher growth rates (see Fig. S2 in the supplemental material). However, since there does not appear to be a smooth inverse relationship between CRP abundance and growth rate, it is unlikely that the CRP concentration functions as a sensor of changes in the cellular growth rate for PguaB.
In a complementary experiment, we asked whether artificially increasing the intracellular concentration of CRP can influence GRDC of PguaB. To do this, we overexpressed the crp gene from its native promoter (plasmid pDCRP) and, separately, from the constitutive bla promoter (plasmid pHA7) in a Δcrp strain and measured the activity of PguaB (−133 to +36) over a range of different growth rates. (Unlike the native crp promoter, the bla promoter is not subject to feedback regulation by CRP or to downregulation in the presence of glucose [36, 54].) PguaB (−133 to +36) activity was also measured in cells containing a plasmid that lacked the crp gene (pDU9). Consistent with our previous observations, positive GRDC of PguaB was not observed in the absence of crp (Fig. (Fig.5A).5A). Overexpression of crp from pDCRP resulted in the restoration of normal GRDC to PguaB (compare Fig. 5B and D). Thus, higher-than-normal levels of CRP do not result in repression of guaB promoter activity relative to that observed in wild-type cells. Placing CRP under the control of the bla promoter also conferred GRDC on PguaB, although the observed slope was steeper than that for cells containing pDCRP (compare Fig. 5C and D). These results demonstrate that constitutively increased expression of crp does not result in constitutive repression of PguaB activity.
Previous studies indicated that the addition of increasing concentrations of cAMP to the growth medium results in progressively increased CRP-mediated repression or activation of target promoters (20, 43, 62). This indicates that the intracellular concentration of cAMP can be manipulated by altering the amount of cAMP added to the medium. Various experimental observations suggest that an exogenous cAMP concentration of 0.5 to 1.0 mM restores intracellular cAMP to a physiologically functional level in adenylate cyclase-deficient, cAMP phosphodiesterase-proficient (i.e., ΔcyaA cpd+) strains, whereas addition of 5 to 10 mM cAMP to the culture medium results in higher-than-normal intracellular concentrations of cAMP (43, 82). Therefore, to examine a possible role for the intracellular cAMP concentration in influencing regulation of PguaB, we measured the activity of PguaB (−133 to +36) in response to the growth rate in a ΔcyaA strain growing in the absence and presence of 5 mM cAMP.
As expected, the presence of the ΔcyaA allele resulted in an increase in PguaB activity at low growth rates relative to that in the wild-type strain background, and this abolished the positive correlation between PguaB activity and the growth rate (Fig. (Fig.6A).6A). Deletion of the cyaA gene also caused a reduction in the growth rate, as observed previously (1, 15, 22, 55). The addition of 5 mM cAMP to media supporting the lowest growth rates caused ΔcyaA bacteria to grow faster than those in cAMP-free media, whereas addition to media supporting the highest growth rates resulted in a decrease in the growth rate relative to those in cAMP-free media (similar observations have been reported previously [38, 43, 66, 82]). This gave rise to a more restricted range of growth rates (Fig. (Fig.6B).6B). Despite the decreased range of growth rates, it is apparent that the relationship between the growth rate and PguaB activity is similar to that observed in wild-type cells growing in the absence of added cAMP (compare Fig. Fig.6B6B with Fig. Fig.5D).5D). This observation suggests that although the CRP-cAMP complex is required for repression of PguaB and for GRDC, variation in the intracellular concentration of cAMP in response to the prevailing carbon source is unlikely to dictate GRDC of this promoter.
cAMP did not restore normal GRDC to the PguaB (−133 to +36, G-122C) promoter, which lacked a functional CRP site, nor did it result in repression of this promoter at low growth rates (compare Fig. 6C and D). Plots of relative activities of the wild-type and mutant guaB promoters in response to cAMP clearly show that the mutant promoter is essentially unresponsive to cAMP (compare Fig. S3A and B in the supplemental material). These results confirm that the cAMP-dependent effects observed at PguaB (−133 to +36) are mediated by CRP binding to the PguaB CRP site.
Disruption of CRP binding to PguaB results in derepression of PguaB activity in media supporting low growth rates and concomitantly abolishes GRDC. To determine whether CRP-cAMP is able to regulate PguaB in the absence of other factors, transcription from PguaB (−133 to +36) was measured in vitro, in the presence and absence of purified CRP-cAMP. Results from multiple-round transcription assays showed that PguaB (−133 to +36) activity was not influenced by CRP-cAMP in the absence of other regulatory factors (Fig. (Fig.7).7). Varying the concentration of the transcription-initiating nucleotide (GTP), salt (KCl), or CRP did not enhance the responsiveness of PguaB to CRP in multiple- or single-round reactions (results not shown). In control experiments, CRP-cAMP was able to activate transcription ~10-fold from the CRP-dependent CC(−41.5) promoter under the same conditions (Fig. (Fig.7).7). The impotence of CRP-cAMP in the in vitro system suggests that an additional factor is required for CRP-dependent downregulation of PguaB.
A putative binding site for the regulatory protein PurR overlaps the guaB promoter (Fig. (Fig.1)1) (31, 48). Moreover, the activity of the guaB gene product, IMP dehydrogenase, in bacteria growing under conditions of purine repression is threefold higher for strains harboring a null mutation in the gene encoding the PurR repressor (48). To assess the role of the PurR repressor in GRDC of the guaB promoter, we first confirmed that PguaB activity is repressed by this regulator. Thus, purine-mediated repression of PguaB (−253 to +36), PguaB (−37 to +36), and PguaB (−253 to +10) was assessed in vivo. All three promoters were shown to be repressed ~50% in the presence of high levels of exogenously added guanine, and this repression was entirely dependent on the presence of a functional purR gene (Fig. 8A and B). This result confirms that transcription of the guaBA operon is repressed by PurR and is consistent with the predicted location of the PurR binding site. The activity of PguaB (−253 to +36) in response to the growth rate was then measured in the purR mutant background. The results showed that there was no significant difference in GRDC of PguaB in the presence or absence of PurR (Fig. 8C and D).
In this study, we have demonstrated that the CRP-cAMP complex binds to a site centered at position −117.5 relative to the PguaB transcription start site, and we have shown that this interaction serves to downregulate PguaB. Moreover, the degree of CRP-mediated repression progressively increases as the bacterial growth rate decreases. Thus, in a crp or cya mutant strain or a strain in which the PguaB CRP site has been inactivated, PguaB exhibits a marked increase in activity at low growth rates. Since glucose supports relatively high growth rates, these results are consistent with a previous transcriptomic analysis in which it was observed that guaB transcription is subject to “CRP-dependent glucose activation” (26). However, our observation is not in agreement with a previous proposal that CRP activates PguaB (35). The latter proposal was based on nutrient downshift experiments in which cells were transferred, after being washed, from a complex nutrient-rich medium to a glycerol-based medium containing Casamino Acids. In these experiments, the authors observed a rapid but short-lived increase in PguaB activity following the downshift. This did not occur in a Δcrp strain. These experiments are difficult to interpret because the nutrient-rich medium also contained a source of purines, so cells had to readjust not only to a change in carbon source but also to purine availability. To avoid the possibility that purine repression would complicate the interpretation of our results, we avoided the use of media containing purines in examining the effects of media supporting different growth rates on PguaB activity. In addition, our measurements were performed during steady-state growth of bacteria.
Since the supply of guanine nucleotides must satisfy the demand for rRNA and tRNA biosyntheses, which are themselves subject to very tight growth rate-dependent control and constitute as much as 85% of the total RNA synthesized at high growth rates, it is not surprising that transcription of the guaBA operon should be subject to a GRDC mechanism (10, 16, 33). However, the observed inverse relationship between CRP-mediated repression of PguaB and the bacterial growth rate prompted us to reexamine whether PguaB was in fact regulated according to the carbon source rather than as a function of the growth rate. Promoters that are subject to GRDC exhibit the same activity in cells growing at the same growth rate, irrespective of the nature of the carbon source. Fig. S4 in the supplemental material shows the activity of PguaB in the presence of different carbon sources. Since some of the carbon sources used gave rise to similar growth rates and PguaB activities, it is clear that the activity of this promoter is influenced by the growth rate rather than the carbon source.
Our results also suggest that E. coli makes use of a mechanism that is distinct from that acting at stable RNA promoters to confer growth rate dependence on PguaB. Unlike the situation at rRNA and tRNA promoters, the mechanism at PguaB employs CRP and requires the participation of the UP element, thereby implying a role for αCTD in this process (33). However, our results do not suggest that GRDC of PguaB is dictated by the intracellular concentration of the CRP-cAMP complex. Thus, manipulating the intracellular cAMP or CRP level so it remains constant at different growth rates does not abolish GRDC. This observation suggests the involvement of another factor that dictates GRDC of PguaB. Furthermore, our in vitro results also suggest that at least one other factor is required for CRP-mediated repression of PguaB. This is consistent with a model in which CRP-mediated repression and GRDC are facets of the same mechanism, which requires at least one other regulatory factor.
How might binding of the CRP-cAMP complex to a site far upstream of the PguaB core promoter region repress transcription? There are no known examples where CRP, acting alone or in concert with an UP element, modulates transcription from a site located more than 100 bp upstream of the transcription start site. However, there are examples where CRP bound at such distances has been shown to contact RNAP and to regulate transcription at promoters that require the participation of additional regulatory proteins (i.e., class III CRP-dependent promoters). For example, CRP can contact αCTD when bound at position −102.5/−103.5 at synthetic class III promoters (8, 72). Moreover, at acsP2, a naturally occurring class III promoter, the promoter-distal CRP site (CRP II) is centered at position −122.5, and CRP bound to this site can interact with αCTD (6). However, the CRP site at PguaB is located on the opposite face of the DNA helix in comparison to the CRP sites at these class III promoters. Therefore, it is not clear that CRP would be in a position to interact with RNAP at PguaB. Moreover, unlike the examples discussed above, the role of CRP at PguaB is to downregulate transcription. One possibility is that CRP (with the aid of an unknown regulatory factor) competes with the UP element for αCTD, thereby recruiting αCTD to a location that is unfavorable for transcription from PguaB. However, we demonstrated that GRDC of PguaB was not altered significantly in strains producing AR1-defective CRP, and an alanine scanning analysis of αCTD did not implicate the AR1 contact site on αCTD (i.e., the 287 determinant) in regulation of this promoter (32, 63). In addition, the AR1 substitution exerted a repressive effect on PguaB, something that would not be expected if the AR1-αCTD interaction was required to downregulate PguaB. Furthermore, displacement of the CRP site from its native position by one turn of the DNA helix in either direction completely abolished GRDC at PguaB. At many promoters where CRP binding to upstream sites serves to regulate transcription by contacting αCTD, displacement of the CRP site by one helical turn from its optimal position does not completely abolish CRP activity (8, 12, 24, 44, 72). Thus, it is unlikely that direct interactions between CRP and RNAP play a role in GRDC of PguaB.
An alternative possibility is that the missing factor contacts RNAP at PguaB in a manner that requires CRP. For example, it has been demonstrated that at some CRP-regulated promoters, CRP can participate in cooperative interactions with other regulatory proteins that activate or repress transcription (14, 53, 79). Thus, at PguaB, CRP could serve as a nucleation site for the assembly of a complex of transcription factors on the DNA, or it could allow remodeling of such an assembly that sterically blocks access to the UP element by αCTD. Steric hindrance of UP element utilization as a regulatory mechanism has been observed previously for LexA at the PrstA promoter of phage CTXΦ (58). Our results suggest that such a mechanism would not be dependent upon interactions between AR1 and the other regulatory protein(s), as observed for other promoters (47, 79). The previous observation that maximum guaB promoter activity can be achieved in the absence of sequences located upstream of position −59 is inconsistent with a mechanism whereby CRP occludes an adjacent or overlapping binding site for a transcription factor that activates PguaB (33).
Putative binding sites for the regulatory proteins DnaA and PurR have been identified within the core region of the guaB promoter (16, 31, 74). PurR mediates the repression of genes required for the biosynthesis of purines in the presence of the corepressors guanine and hypoxanthine (49). It has been shown that the presence of guanine in the culture medium results in similar degrees of repression of PguaB at all growth rates, and a purR mutation results in increased activity of the guaB gene product, IMP dehydrogenase, suggesting that PurR regulates PguaB (16, 48). However, while confirming that PurR is responsible for guanine-mediated repression of PguaB, our results rule out a role for PurR in GRDC at this promoter. Regarding DnaA, a second binding site for this protein, located in the guaB coding region, is also required for DnaA-mediated repression of PguaB (73, 74). Since the promoters used for assessment of GRDC in the current study contained only the upstream DnaA binding site, which does not serve to repress PguaB in the absence of the downstream site, this argues against a role for DnaA in GRDC. Thus, the missing factor is unlikely to be PurR or DnaA.
The expression of many genes, including those encoding components of the translation apparatus, the RNAP-associated protein RapA, and Dam methylase, is subject to positive GRDC, a phenomenon characterized by an increase in the gene product/bacterial mass ratio as the growth rate increases (13, 19, 57, 60, 78). However, the responsiveness of individual growth rate-regulated promoters to the growth rate varies. Accordingly, there appear to be diverse mechanisms for implementing GRDC, and in some cases it is exerted at a posttranscriptional level (5, 13, 19, 55, 57, 65). However, a complete understanding of the mechanism of GRDC at any single promoter remains elusive. In conclusion, our results implicate a role for the CRP-cAMP complex in GRDC of PguaB and highlight the diverse range of mechanisms that can serve to modulate gene expression in response to the growth rate.
We are indebted to W. Ross, T. Gaal, H. Murray, and R. L. Gourse (University of Wisconsin-Madison) and to D. Browning, D. Grainger, and D. Lee (University of Birmingham) for purified CRP, strains, plasmids, and helpful discussions. We also thank H. Aiba (Nagoya University) for anti-CRP antiserum, strains, and plasmids.
This work was supported by a project grant awarded by the Wellcome Trust (grant 073917) and by a family Ph.D. sponsorship awarded to S.I.H., kindly provided by S. M. Husnain.
Published ahead of print on 24 July 2009.
†Supplemental material for this article may be found at http://jb.asm.org/.