The UP Element Is Necessary but Not Sufficient for Growth Rate-Dependent Control of the Escherichia coli guaB Promoter

Abstract

The Escherichia coli guaB promoter (P_guaB) regulates the transcription of two genes, guaB and guaA, that are required for de novo synthesis of GMP, a precursor for the synthesis of guanine nucleoside triphosphates. The activity of P_guaB is subject to growth rate-dependent control (GRDC). Here we show that the A+T-rich sequence located between positions −59 and −38 relative to the guaB transcription start site stimulates transcription from P_guaB ∼8- to 10-fold and, in common with other UP elements, requires the C-terminal domain of the RNA polymerase α subunit for activity. Like the rrnB P1 UP element, the P_guaB UP element contains two independently acting subsites located at positions −59 to −47 and −46 to −38 and can stimulate transcription when placed upstream of the lacP1 promoter. We reveal a novel role for the P_guaB UP element by demonstrating that it is required for GRDC. The involvement of the UP element in GRDC also requires the participation of sequences located at least 100 bp upstream of the guaB transcription start site. These sequences are required for down-regulation of P_guaB activity at lower growth rates.

Escherichia coli RNA polymerase (RNAP) is a multisubunit complex consisting of an α subunit homodimer, single β, β′, and ω subunits, and one of several σ subunits (14). The α subunit contains an N-terminal domain (αNTD; residues 8 to 232) and a C-terminal domain (αCTD; residues 249 to 329) (5, 21, 44). αNTD and αCTD are separated by a flexible linker that is at least 13 amino acids long (5, 20, 25). αCTD plays important roles in transcription activation by contacting a number of transcription factors (6, 18) or through interacting with an A+T-rich DNA sequence referred to as the UP element (16).

UP elements stimulate transcription by recruiting αCTD to DNA (32, 43). The best-studied UP element spans the region from −59 to −38 at the rrnB P1 promoter (16). The rrnB P1 UP element contains two domains, a ∼9-bp proximal UP element subsite located at positions −46 to −38 and a ∼13-bp distal UP element subsite located at positions −59 to −47. Each subsite functions independently by recruiting one αCTD (11). The amino acid residues in αCTD that comprise the 265 determinant, including residues V264, R265, N268, N294, G296, K298, and S299, interact with UP element and A-tract sequences by contacting the DNA backbone, and R265 makes additional contacts with bases in the minor groove (4, 32, 43). These residues are the most important for rrnB P1 UP element utilization both in vivo and in vitro (13, 35). In the absence of a strong distal UP element subsite, αCTD bound to a consensus or near-consensus proximal UP element subsite also makes productive contacts with region 4.2 of σ⁷⁰ that serve to stimulate transcription (7, 35).

The E. coli guaB promoter (P_guaB) regulates the transcription of two genes, guaB and guaA, which together constitute the guaBA operon. The guaB and guaA genes encode IMP dehydrogenase and GMP synthetase, respectively, and are required for the biosynthesis of GMP from IMP (22, 42). P_guaB is regulated by the cyclic AMP (cAMP) receptor protein (CRP), and CRP binds to a ∼300-bp DNA fragment containing P_guaB that includes a predicted CRP site centered at position −117.5 relative to the guaB transcription start site (19). P_guaB also contains putative binding sites for PurR and DnaA overlapping the core promoter region (23, 41). In addition, an A+T-rich sequence located between positions −59 and −38 at P_guaB matches the consensus 22-bp UP element sequence [5′-nnAAA(A/T)(A/T)T(A/T)TTTTnnAAAAnnn-3′; n represents a nonconserved position]) at 13 out of 15 conserved positions (Fig. 1), but its role as an UP element has not been confirmed (10, 19).

FIG. 1.

Comparison of A+T-rich sequences used in this work with the consensus sequence for UP elements. Consensus UP element sequences and the putative P_guaB UP element (positions −59 to −38 relative to the guaB transcription start site) are aligned with the 22-bp general consensus sequence for UP elements (10). Nonconserved positions are indicated by a lowercase “n.” Mismatches to the general consensus UP element sequence are shown in boldface. The degree of the match to the general consensus is shown for each sequence. The consensus 4541 UP element was made by combining optimal proximal and distal UP element subsites (11). The consensus 4192 UP element was selected by a SELEX procedure (10). The predicted P_guaB proximal UP element subsite contains P_guaB sequences from position −46 to −38, whereas the predicted P_guaB distal UP element subsite contains sequences from −59 to −47. The SUB sequence is nonspecific for αCTD and contributes minimally to transcription (10, 30).

The term growth rate-dependent control (GRDC) is used to refer to situations where the abundance of a particular transcript or protein product represents an increasing fraction of the cell mass at higher growth rates. For example, in E. coli, the synthesis rates of rRNA and tRNA increase in proportion to the square of the growth rate, with the result that stable RNA accounts for an increased fraction of the cellular mass with increasing growth rate (9, 15, 27). One consequence of GRDC of stable RNA synthesis is that the consumption of nucleoside triphosphates also increases in a growth rate-dependent manner (28). Because GMP is a precursor for the synthesis of GTP, it follows that the rate of synthesis of GMP should be regulated according to the growth rate. Consistent with this, it has been shown that P_guaB activity is subject to GRDC (8). The best-characterized promoter subject to GRDC that also contains an UP element is the rrnB P1 promoter, although the UP element plays no role in GRDC of this promoter (2, 28). In this work, we show that the A+T-rich sequence located upstream of the P_guaB −35 region does indeed function as an UP element and that, like the rrnB P1 UP element, it contains promoter-proximal and -distal subsites that are able to activate transcription independently. However, we demonstrate that, in contrast to the situation at the rrnB P1 promoter, the P_guaB UP element is required for GRDC of P_guaB and this regulation requires the participation of sequences further upstream.

MATERIALS AND METHODS

Strains.

The bacterial strains and plasmids used in this study are listed in Table 1. For promoter measurements in vivo, strains containing single-copy promoter-lacZ transcriptional fusions were employed. All transcriptional fusions were carried on λ prophages and were constructed using a system based on λimm21 (30, 40). All P_guaB-lacZ fusions were made in the VH1000 genetic background and had downstream endpoints at +36 with respect to the guaB transcription start site. All lacP1-lacZ transcriptional fusions were made in strain NK5031. Apart from strain RLG4288, lacP1 derivatives each contain the core lacP1 promoter (positions −37 to +52) with non-lac sequences located upstream of position −37 relative to the lacP1 transcription start site. Strain NKGLAC contains a lacP1-lacZ fusion in which the predicted P_guaB UP element is located at positions −59 to −38.

TABLE 1.

Bacterial strains and plasmids

Strain or plasmid	Genotype/descriptiona	Promoter designationb	Source or reference
Bacterial strains
NK5031	F′ lacZΔM5262 supF Nalr		17
VH1000	MG1655 lacZ lacI pyrE+		12
NKGLAC	NK5031/λ [SUB(−66 to −60)-PguaB(−59 to −38)-lacP1(−37 to +52)]-lacZ		This study
RLG4208	NK5031/λ [rrnB P1(−66 to −60)-4192(−59 to −38)-lacP1(−37 to +52)]-lacZ		10
RLG4282	NK5031/λ [rrnB P1(−88 to −38, Δ72)-lacP1(−37 to +52)]-lacZ		31
RLG4288	NK5031/λ [SUB(−58 to −41)-lacP1(−40 to +52)]-lacZ	Core lacP1	31
VH1000G-253	VH1000/λ [PguaB(−253 to +36)]-lacZ	PguaB(−253 to +36)	This study
VH1000G-133	VH1000/λ [PguaB(−133 to +36)]-lacZ	PguaB(−133 to +36)	This study
VH1000G-133SUB	VH1000/λ [PguaB(−133 to −60)-SUB(−59 to −38)-PguaB(−37 to +36)]-lacZ	PguaB(−133 to +36) (no UP)	This study
VH1000G-100	VH1000/λ [PguaB(−100 to +36)]-lacZ	PguaB(−100 to +36)	This study
VH1000G-84	VH1000/λ [PguaB(−84 to +36)]-lacZ	PguaB(−84 to +36)	This study
VH1000G-69	VH1000/λ [PguaB(−69 to +36)]-lacZ	PguaB(−69 to +36)	This study
VH1000G-59	VH1000/λ [SUB(−66 to −60)-PguaB(−59 to +36)]-lacZ	PguaB(−59 to +36)	This study
VH1000G-PROX	VH1000/λ [SUB(−66 to −47)-PguaB(−46 to +36)]-lacZ		This study
VH1000G-DIS	VH1000/λ [SUB(−66 to −60)-PguaB(−59 to −47)-SUB(−46 to −38)-PguaB(−37 to +36)]-lacZ		This study
VH1000G-CONS	VH1000/λ [4541(−59 to −38)-PguaB(−37 to +36)]-lacZ		This study
VH1000G-37	VH1000/λ [SUB(−59 to −38)-PguaB(−37 to +36)]-lacZ	Core PguaB	This study
Plasmids
pBluescript II KS	E. coli-specific cloning vector (Apr)		1
pMSB1	Derivative of promoter cloning plasmid, pRS1553, for construction of lacZ fusions (Apr)		30
pRLG770	Derivative of pKM2 for in vitro transcription reactions		36

^aAp^r, ampicillin resistance. The architecture of each promoter is shown in brackets. The SUB sequence is a DNA sequence that is nonspecific for αCTD. 4192 is the consensus 4192 UP element (10); 4541 is the consensus 4541 UP element (11). For SUB and UP element sequences, see Fig. 1. The SUB sequence inserted at positions −66 to −60 (5′-GACTGCA-3′) separates an UP element from the upstream A+T-rich EcoRI site. At other promoters, the corresponding region of SUB is used to replace the UP element sequences shown in Fig. 1. Δ72 indicates a deletion of an A:T base pair at position −72 at rrnB P1, which results in loss of binding and activation by the factor for inversion stimulation (36).

^bDesignations of promoter derivatives referred to in the text are given where appropriate.

Media and growth conditions.

Bacteria were grown at 37°C. NK5031 strain derivatives, each containing a lacP1-lacZ transcriptional fusion, were grown in Luria-Bertani medium. VH1000 strain derivatives, each containing a P_guaB-lacZ transcriptional fusion, were grown in M9 minimal medium supplemented with 0.2% (wt/vol) glucose and 5 μg/ml thiamine. Strains containing plasmids encoding wild-type (WT) α or mutant α derivatives were grown in M9 minimal medium with 0.4% (wt/vol) glucose, 5 μg/ml thiamine, 0.8% (wt/vol) Casamino Acids, and 100 μg/ml ampicillin. To achieve different cellular growth rates, cells were grown in M9 minimal medium containing 5 μg/ml thiamine and one of the following (in order of increasing growth rate supported): 0.4% (wt/vol) glycerol, 0.4% succinate plus 20 amino acids, 0.4% (wt/vol) glycerol plus 20 amino acids, 0.4% (wt/vol) glycerol plus 1% (wt/vol) Casamino Acids, 0.4% (wt/vol) glycerol plus 2% (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 (“20 amino acids” comprised each amino acid at a final concentration of 20 μg/ml).

Recombinant DNA techniques.

Standard methods for plasmid isolation and DNA manipulation were used throughout (37). All promoter derivatives were made by PCR and contained an upstream EcoRI restriction site and a downstream HindIII restriction site. PCR products were digested with EcoRI and HindIII, ligated to pMSB1 for use in construction of single-copy lacZ fusions (30), and also ligated to pRLG770 for in vitro transcription assays. All amplified sequences were checked by DNA sequencing.

Measurement of transcription in vivo.

Dense starter cultures of bacteria containing single-copy promoter-lacZ fusions were diluted 100-fold into the same growth medium and grown at 37°C to an optical density at 600 nm (OD₆₀₀) of 0.3 to 0.5. For GRDC experiments, starter cultures were grown overnight in medium supporting the lowest growth rate (i.e., using glycerol as the sole carbon source). Strains containing plasmids encoding WT α and mutant α derivatives were inoculated from fresh colonies into minimal medium at an OD₆₀₀ of ∼0.007 and grown at 37°C to an OD₆₀₀ of ∼0.335 to 0.45 in the presence of 100 μg/ml ampicillin. Cells were permeabilized using chloroform-sodium dodecyl sulfate, and the β-galactosidase activity was determined using a published protocol (26). For GRDC experiments, cells were disrupted by sonication, and the β-galactosidase activity was normalized to the activity corresponding to a growth rate of 1 doubling per h, which was calculated from a linear regression curve fitted to a plot of β-galactosidase activity versus growth rate.

Measurement of transcription in vitro.

Multiple-round transcription reactions were performed at 30°C for 20 min using a previously described method (36). Each 25-μl reaction mixture contained 40 mM Tris-acetate (pH 7.9), 150 mM KCl, 10 mM MgCl₂, 1 mM dithiothreitol, 100 μg/ml acetylated bovine serum albumin (Promega), 200 μM ATP, 200 μM GTP, 200 μM CTP, 10 μM UTP, and 5 μCi of [α-³²P]UTP (800 Ci/mmol; Perkin-Elmer). Each reaction mixture also contained 0.2 nM supercoiled template DNA (pRLG770 derivative) and either a commercially available WT E. coli RNAP holoenzyme (Epicenter) or E. coli core RNAP purified as described elsewhere (E. Campbell, S. Darst, S. I. Husnain, B. Nickels, M. S. Thomas, K.-A. Twist, and L. Westblade, unpublished data) and subsequently reconstituted with σ⁷⁰. The latter form of RNAP contained either WT α or a derivative lacking αCTD (RNAPΔαCTD), and RNAP activity was standardized as described previously (Campbell et al., unpublished). Reactions were initiated with RNAP and terminated with 25 μl of stop solution (95% formamide, 20 mM EDTA, 0.05% bromophenol blue, 0.05% xylene cyanol). Samples, containing equal amounts of template DNA, were fractionated on a 5.5% acrylamide gel containing 7 M urea, and transcript abundance was quantified using a FujiFilm FLA3000 phosphorimager. The amount of transcript was determined by deducting the background activity in the lane from the value for the transcript abundance.

DNase I footprinting.

DNase I footprinting was performed according to a published method (38). Derivatives of pBluescript II KS containing EcoRI-XhoI DNA fragments extending from position −133 to +36 of the guaB promoter were purified following electrophoresis in 6% acrylamide gels (25), labeled at the downstream (XhoI) end with [γ-³²P]ATP (6,000 Ci/mmol; Perkin-Elmer) using T4 polynucleotide kinase, and purified using QIAQuick columns (Qiagen). Labeled DNA fragment (4 nM) was incubated at room temperature for 30 min in a buffer containing 20 mM HEPES (pH 8.0), 5 mM MgCl₂, 50 mM potassium glutamate, 1 mM dithiothreitol, 500 μg/ml acetylated bovine serum albumin, and 7% glycerol in the presence or absence of either 55 nM WT reconstituted RNAP or 119 nM reconstituted RNAPΔαCTD, in a final volume of 20 μl, after which DNA fragments were digested with DNase I. Reactions were stopped using a buffer containing 0.3 M sodium acetate (pH 5.2) and 10 mM EDTA. Samples were purified by phenol-chloroform extraction and ethanol precipitation, and DNA fragments were separated in a 6% acrylamide-7 M urea sequencing gel. A Maxam-Gilbert G+A sequencing ladder was run alongside. Footprints were visualized using a FujiFilm FLA3000 phosphorimager.

RESULTS

Activation of P_guaB in vivo by sequences located upstream of the core promoter elements.

To analyze the contribution to transcription of sequences present upstream of P_guaB, promoter derivatives with different upstream endpoints were used to generate transcriptional fusions to the lacZ gene and introduced into the E. coli chromosome as a single copy. The full-length WT P_guaB derivative used in this study contained promoter sequences extending upstream to position −253 relative to the guaB transcription start site. This promoter and a promoter derivative containing sequences up to −133 both harbored the putative CRP site. In the other promoter derivatives, the upstream deletion removed the putative CRP site. All promoter derivatives had a downstream endpoint located at +36 (Fig. 2A). The promoter activity was determined by measuring the β-galactosidase activity in exponentially growing cells.

FIG. 2.

Contributions of sequences upstream of the guaB promoter to transcription. (A) The WT guaB promoter (−253 to +36) and derivatives each having a different upstream endpoint were constructed. Deletion endpoints are indicated relative to the guaB transcription start site. The predicted CRP binding site centered at position −117.5, the putative P_guaB UP element (designated “UP”), the P_guaB −35 and −10 hexamers, and the guaB transcription start site are shown (diagram not to scale). P_guaB(−37 to +36) is referred to as core P_guaB and contains the SUB sequence from positions −59 to −38 (Fig. 1). (B) The β-galactosidase activity (in Miller units) specified by each promoter-lacZ fusion in strain VH1000 growing logarithmically in M9-glucose minimal medium is expressed relative to the activity of P_guaB(−253 to +36) (activity, 3,003 Miller units; relative activity, 1). (C) Multiple-round in vitro transcription assays were performed using a subset of guaB promoter derivatives with different upstream endpoints cloned into plasmid pRLG770, in the presence of 10 nM RNAP. The P_guaB-directed transcript produced from this plasmid contains ∼180 nucleotides (RNA-I contains ∼110 nucleotides). Each promoter activity was determined by measuring the transcript abundance and is expressed relative to the activity of P_guaB(−253 to +36), set at 1. In panels B and C, values are means and standard deviations from three independent experiments.

The results from the deletion analysis revealed that removal of the sequence from position −253 to −60 does not lead to any significant changes in the efficiency of transcription from P_guaB (Fig. 2B). This suggests that sequences upstream of position −60 do not contribute significantly to transcription from P_guaB in cells growing under the conditions employed. Transcription was also measured from a P_guaB derivative with an upstream endpoint at position −37, which is referred to as the core guaB promoter (see the legend to Fig. 2). The core guaB promoter contains a “SUB” sequence in place of the putative P_guaB UP element at positions −59 to −38 (Fig. 1). The SUB sequence has been shown to minimize the contribution to transcription from the recruitment of αCTD to DNA (10, 30). Comparison of the activity of P_guaB(−59 to +36) with core P_guaB indicated that the putative P_guaB UP element stimulates the core promoter approximately eightfold in vivo (Fig. 2B).

Activation of P_guaB in vitro by sequences located upstream of the core promoter elements.

To determine whether upstream guaB promoter sequences stimulate transcription in the absence of additional regulatory factors, multiple-round in vitro transcription reactions were performed using a subset of the promoter derivatives employed for the in vivo study (see above). Reactions were initiated with 10 nM RNAP. The results showed that, in the absence of other factors, sequences located upstream of position −59 do not contribute significantly to the transcriptional activity of P_guaB. In contrast, deletion of the putative P_guaB UP element gave rise to a ∼6-fold decrease in transcription activity compared to that of the full-length guaB promoter (Fig. 2C). These results are in good agreement with the findings from the in vivo analysis (see above) and demonstrate that the putative P_guaB UP element can function in the absence of additional regulatory factors.

The putative P_guaB UP element contains independently acting subsites.

We measured transcription both in vivo and in vitro from P_guaB(−59 to +36), which contained the putative P_guaB UP element sequence, and P_guaB derivatives containing only the predicted P_guaB proximal UP element subsite, only the predicted P_guaB distal UP element subsite, or a consensus UP element (Fig. 1). The locations of the P_guaB UP element subsites used here correspond to the proximal and distal UP element subsites defined in a previous study (11). The consensus UP element used in this analysis (the 4541 UP element [Fig. 1]) has been described previously (11). The results were compared to the transcription activity of core P_guaB.

The results showed that the contributions of each tested sequence to transcription in vivo and in vitro are comparable (Fig. 3A and B). Whereas the putative P_guaB UP element afforded an ∼8- to 10-fold increase in transcription from P_guaB, the putative P_guaB proximal UP element subsite stimulated transcription ∼5- to 6-fold and the putative P_guaB distal UP element subsite afforded a ∼3-fold increase in transcription. These results demonstrate that the putative P_guaB UP element contains subsites that can function independently in stimulating transcription from P_guaB and that together they function as coactivators of transcription. The consensus UP element stimulated transcription from P_guaB ∼11-fold, i.e., similar to the level afforded by the putative P_guaB UP element.

FIG. 3.

Contribution of subsites within the P_guaB A+T-rich sequence to transcription. guaB promoter derivatives containing the consensus 4541 UP element located at positions −59 to −38, the complete 22-bp P_guaB A+T-rich element located at −59 to −38 (full), the predicted P_guaB proximal UP element subsite (positions −46 to −38), or the predicted P_guaB distal UP element subsite (positions −59 to −47), or lacking an UP element (core P_guaB), were constructed. Core P_guaB contains a SUB sequence from position −59 to −38. P_guaB containing the proximal UP element subsite also contains a SUB sequence from position −59 to −47, and P_guaB containing the distal UP element subsite contains a SUB sequence from position −46 to −38. (A) The β-galactosidase activity (in Miller units) specified by each promoter-lacZ fusion in strain VH1000 growing logarithmically in M9-glucose minimal medium is expressed relative to the activity of core P_guaB (activity, 385 Miller units, relative activity, 1). (B) Multiple-round in vitro transcription assays were performed using each guaB promoter construct in plasmid pRLG770 in the presence of 5 nM RNAP. The activity of each promoter is expressed relative to that of core P_guaB, set at 1. In both panels, values are means and standard deviations from three independent experiments.

Activation of the lac promoter in vivo by the putative P_guaB UP element.

The rrnB P1 UP element is capable of stimulating transcription from other promoter sequences (31). To determine whether this also holds true for the putative P_guaB UP element, transcription was measured in vivo from the core lac promoter (core lacP1) containing the putative P_guaB UP element located at positions −59 to −38. To compare the strength of activation due to the putative P_guaB UP element with that from other UP element sequences, transcription from core lacP1 fused to either the rrnB P1 UP element or a consensus UP element was also quantified. The consensus UP element used here (the 4192 UP element [Fig. 1]) has been described previously (10). The results were compared to the transcription activity of the core lac promoter (i.e., lacP1 without an UP element). The results show that the putative P_guaB UP element stimulated transcription from core lacP1 ∼23-fold, i.e., at a much higher level than stimulation of the guaB promoter (Fig. 4). By comparison, the rrnB P1 UP element stimulated transcription from the core lac promoter ∼30-fold, whereas the consensus UP element afforded an ∼80-fold increase in promoter activity.

FIG. 4.

Activation of the lacP1 promoter by the P_guaB A+T-rich element. lacP promoter derivatives contain WT lacP1 sequences from −37 to + 52 and one of the following: the consensus 4192 UP element (“Consensus”), the rrnB P1 UP element, the P_guaB UP element, or no UP element, as indicated. UP element sequences are positioned from −59 to −38 relative to the lacP1 transcription start site. The core lac promoter contains lac promoter sequences as far upstream as position −40 and a SUB sequence from positions −58 to −41 (31). The β-galactosidase activity (in Miller units) specified by each promoter-lacZ fusion is the mean (and standard deviation) from three independent experiments and is expressed as a percentage of the activity of the core lac promoter (activity, 36 Miller units; relative activity, 1).

Role of αCTD in transcription from P_guaB in vivo.

To investigate whether αCTD plays a role in transcription at P_guaB, transcription from P_guaB(−59 to +36) and core P_guaB was measured in merodiploid cells containing the WT rpoA allele on the chromosome and a plasmid encoding either WT α or a mutant derivative harboring an alanine substitution at position 265 (α265A). The results show that the activity of P_guaB(−59 to +36) did not differ significantly in the presence or absence of α265A (see Fig. S1A in the supplemental material). However, in the presence of α265A, transcription from core P_guaB increased ∼2.9-fold over that for cells containing only WT α. This can be explained by the rationale that a decrease in the activity of the putative UP element in the presence of α265A leads to a compensatory increase in transcription from core P_guaB. The effect of the mutant α on UP element function was therefore quantified as described previously (13, 24, 33). Thus, in the presence of α265A, the stimulation of P_guaB due to the putative UP element was reduced by a factor of ∼2.5 (from 9.5-fold to 3.6-fold) (see Fig. S1B in the supplemental material).

Role of αCTD in transcription from P_guaB in vitro.

To demonstrate that αCTD is required for stimulation of P_guaB by the A+T-rich element, transcription was also measured in vitro from P_guaB(−59 to +36) and from core P_guaB in the presence of the RNAP holoenzyme containing WT α or α lacking αCTD (RNAPΔαCTD). The results show that transcription from P_guaB(−59 to +36) decreased ∼10-fold in the presence of RNAPΔαCTD in comparision to the activity measured in the presence of WT RNAP. On the other hand, the activity of core P_guaB did not differ significantly in the presence of either RNAP, and in each case the promoter activity was comparable to the activity of P_guaB(−59 to +36) in the presence of RNAPΔαCTD (Fig. 5A). This indicates that the P_guaB A+T-rich element requires αCTD to function in transcription activation.

FIG. 5.

Role of αCTD in transcription from P_guaB in vitro. (A) Multiple-round transcription reactions were performed with the core guaB promoter and P_guaB(−59 to +36) in the presence of either 5 nM WT reconstituted RNAP or 11 nM reconstituted RNAPΔαCTD. The concentrations of the two RNAPs employed showed comparable activities, as measured by multiple-round transcription from the lacUV5 promoter (data not shown). The activity of each promoter is expressed as a percentage of that of P_guaB(−59 to +36) in the presence of WT RNAP, set at 100%. Each promoter activity is the mean and standard deviation from three independent experiments. (B) DNase I footprinting was performed with the P_guaB(−133 to +36) DNA fragment (WT) and with P_guaB(−133 to +36) (no UP). Experiments were performed in the presence of either WT reconstituted RNAP or RNAPΔαCTD. The region protected by RNAP and the locations of the UP element and SUB sequence are indicated. Nucleotide positions are shown relative to the guaB transcription start site, and lanes containing the G+A ladder are indicated.

DNase I footprinting analysis was employed to probe for αCTD-DNA interactions at P_guaB. The putative P_guaB UP element contains a GC dinucleotide at positions −46 and −45 (Fig. 1) that is susceptible to cleavage by DNase I. In the presence of WT RNAP, cleavage at this position in P_guaB(−133 to +36) is abolished, but this protection is lost in the presence of RNAPΔαCTD, indicating that αCTD interacts with this region of the guaB promoter (Fig. 5B). Replacement of the putative P_guaB UP element with a SUB sequence [i.e., P_guaB(−133 to +36) (no UP)] did not abolish protection of the region corresponding to the proximal UP element subsite by WT RNAP (i.e., the SUB sequence was still bound by αCTD).

The P_guaB UP element is required for GRDC.

GRDC of P_guaB was examined by growing cells harboring single-copy P_guaB-lacZ fusions in media that supported different cellular growth rates and measuring the β-galactosidase activity. The activity of P_guaB(−133 to +36) was found to increase in direct proportion to an increasing growth rate, i.e., a twofold increase in the growth rate was accompanied by a twofold increase in promoter activity (Fig. 6A). As expected, deletion of upstream sequences to position −37 (core P_guaB) resulted in a decrease in guaB promoter activity at all growth rates (Fig. 6B). However, the decrease was progressively more marked at higher growth rates, and this resulted in a complete loss of GRDC. To investigate whether the UP element contributes to GRDC, the UP element in P_guaB(−133 to +36) was replaced by a SUB sequence [i.e., P_guaB(−133 to +36) (no UP)]. This manipulation also resulted in a loss of GRDC (as well as causing a sharp decrease in guaB promoter activity), suggesting that the UP element is required for GRDC (Fig. 6C). To determine whether sequences located upstream of the UP element are also required for GRDC, the activity of P_guaB(−59 to +36) was measured at different cellular growth rates. The results demonstrated that sequences located upstream of the UP element are also required for GRDC (Fig. 6D). Moreover, at the lower growth rates (i.e., μ ≤ 1.0), this promoter derivative is at least 2.3-fold more active than P_guaB(−133 to +36), suggesting that sequences located upstream of the UP element are required for the repression of P_guaB in slowly growing cells. To localize these sequences more precisely, the activity of the P_guaB(−100 to +36) promoter was also measured at different growth rates. The results show that this promoter derivative behaves similarly to P_guaB(−59 to +36) (Fig. 6E), implying that sequences located between −133 and −100 participate in the GRDC of P_guaB.

FIG. 6.

Requirement of the P_guaB UP element for GRDC. The effects of growth rate on the activities of the indicated P_guaB derivatives were determined by growing strain VH1000 harboring fusions of these promoters to lacZ in different media and determining the β-galactosidase activities. The activity of each promoter was normalized to a value of 1.0 at a growth rate of 1 doubling per h. The β-galactosidase activity specified by each promoter (in Miller units) at a growth rate of 1 doubling per h was as follows: P_guaB(−133 to +36), 1,588; P_guaB(−37 to +36), 418; P_guaB(−133 to +36) (no UP), 419; P_guaB(−59 to +36), 3,573; P_guaB(−100 to +36), 3,009. Each data point represents the mean promoter activity and the mean growth rate. Means were calculated using data obtained from three independent experiments.

DISCUSSION

Previously, a 22-bp UP element-like sequence extending from position −59 to −38 relative to the guaB transcription start site was identified by comparison to the consensus sequence for UP elements (10, 19). We demonstrate here that this sequence does function as an UP element. First, this sequence stimulates transcription from P_guaB ∼8- to 10-fold both in vivo and in vitro. Moreover, it can activate other promoters such as lacP1, which it can stimulate at least 20-fold. The variation in transcription enhancement at different promoters is a feature of other UP elements (for example, a consensus UP element stimulates transcription from P_guaB ∼11-fold, from rrnB P1 ∼340-fold, and from lacP1 ∼80-fold in vivo) (11, 31). The lower contribution of the P_guaB UP element to transcription from P_guaB relative to that for lacP1 is likely to be due to differences in promoter architecture that result in a higher activity of the core guaB promoter in comparison to lacP1 (i.e., core P_guaB is ∼10-fold more active than core lacP1). In particular, P_guaB has a consensus −10 sequence that is separated from the −35 region by 17 bp, whereas at lacP1 the −10 sequence (which has a poorer match to the consensus sequence) is separated from the −35 region by an 18-bp spacer.

Second, the P_guaB UP element appears to contain independently functioning subsites that correspond in location to the previously identified proximal and distal subsites in the rrnB P1 and consensus UP elements (11). Accordingly, we assume that each P_guaB UP element subsite functions by contacting a single αCTD. As with the rrnB P1 UP element, the P_guaB-proximal UP element subsite stimulates transcription to a greater degree than the distal subsite (11). It has been shown that αCTD bound to consensus or near-consensus proximal UP element subsites functionally interacts with region 4 of σ⁷⁰ and that such interactions may account for the observation that proximal UP element subsites contribute more to transcription activation than distal UP element subsites (7, 35). Since the P_guaB-proximal UP element subsite matches the consensus sequence, it is possible that αCTD-σ⁷⁰ interactions may be important for P_guaB UP element function.

Third, the activity of the P_guaB UP element requires αCTD. Thus, deletion of αCTD abolishes UP element-dependent transcription from P_guaB in vitro and is accompanied by a loss of protection of the UP element by RNAP. Although protection of non-UP element sequences located upstream of the core guaB promoter by WT RNAP was observed, these interactions do not contribute a strong stimulatory effect on transcription. This observation mirrors findings from a previous study that demonstrated protection of the region between −35 and −70 at the lacUV5 promoter by αCTD, although these sequences do not contain an UP element (34). Incorporation into RNAP of α subunits containing an amino acid substitution in the DNA binding region (α265A) was also shown to decrease UP element-dependent transcription from P_guaB in vivo. However, in this situation, UP element-independent transcription from the core guaB promoter fully compensated for the reduction in UP element function. This suggests that core P_guaB is subject to a feedback mechanism that can counter regulatory inputs acting outside of the core promoter region. Although a compensatory mechanism (involving the initiating nucleotide for transcription) operates under similar conditions at the rrnB P1 promoter, it does not restore promoter activity completely in the presence of α265A (13, 39). This is probably due to the much larger negative effect that impairment of the αCTD-UP element interaction exerts on rrnB P1 activity.

The P_guaB UP element is absolutely required for GRDC of P_guaB. This contrasts with the situation at two other growth rate-regulated promoters (rrnB P1 and the leuV promoter), where GRDC does not require the UP element (2, 3, 29). Our results also show that, in addition to the P_guaB UP element, sequences located between positions −133 and −100 are required for GRDC, and their role is to facilitate the inhibition of P_guaB at lower growth rates. This region contains a sequence exhibiting a good match to the CRP site consensus sequence, suggesting that CRP may participate in GRDC of P_guaB. These observations do not rule out the possibility that an additional factor(s) that binds to the region between −100 and the UP element is involved. Therefore, GRDC of P_guaB may be explained by one of the following two general models. (i) Binding of one or more regulatory proteins to sequences located upstream of, or overlapping, position −100 (for example, CRP) inhibits UP element function. This could occur directly or indirectly, by facilitating occupancy or remodeling of the region upstream of (and overlapping) the UP element by other DNA-binding proteins. (ii) The P_guaB UP element, with or without the participation of other DNA-binding proteins, may alter the orientation of the upstream DNA so as to allow a regulator bound near position −100 to inhibit RNAP, perhaps by bringing it into close proximity to the core promoter region. In both models, DNA occupancy by at least one of the DNA-binding proteins, although not necessarily the factor binding upstream of position −100, would increase at lower growth rates. Work is currently under way to explore these possibilities.