| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 280, Issue 50, 41315-41323, December 16, 2005
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
38 OR 32*
From the Department of Biochemistry and Molecular and Cellular Biology of Plants, Estación Experimental del Zaidín, Consejo Superior de Investigaciones Científicas, Apartado de Correos 419, E-18008 Granada, Spain
Received for publication, May 17, 2005 , and in revised form, September 28, 2005.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
32 or E38 according to the growth phase, with an identical transcriptional start site. To investigate sequence determinants in the interaction between either of the two RNA polymerases and Pm, all possible single mutants between positions -7 and -18 were generated, and the activity in the exponential and stationary phases of the resulting mutant promoters was compared. The results precisely delimited a -10 element between positions -7 and -12 (TAGGCT), which defined a promoter sharing nucleotides with both 38 and 32 consensus. The first two and the last positions of this hexamer were crucial for recognition by both polymerases. Position -10 was the only one specifically recognized by E38, whereas positions -8, -9, and the C-track between positions -14 and -17 were important for specific E32 recognition. Western blots showed that 32 was only detectable in the exponential phase, and 38 appeared in the early stationary phase. In the rpoH mutant KY1429, 38 was already present in the exponential growth phase both free and bound to the RNA polymerase core, in good correlation with the transcription levels found. Pm seems to be optimized for recognition by 32 as an initial response to the addition of effector to the medium and allows binding of the adaptable 38-dependent RNA polymerase in the stationary phase. XylS is always required for Pm transcription. Therefore, the mechanism that controls Pm expression involves specific nucleotide sequences, the abundance of free and core-bound 32 and 38 factors during growth, and the presence of the regulator activated by an effector. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
The Pm promoter seems to be a Class II promoter in which complex interactions occur between the RNA-polymerase and its cognate transcriptional regulator (12). In vivo and in vitro methylation assays of Pm show extensive methylation of T at position -41 in the bottom strand, suggesting the presence of a key distortion point that may favor XylS/RNA polymerase interactions (8). In this connection, it has been shown that XylS contacts residues 291 and 289 of the 
-subunit of RNA polymerase (13). The Pm promoter is unique in that in vivo transcription is mediated by RNA polymerase with different alternative factors. Transcription from the Pm promoter in the early exponential growth phase is mediated by RNA polymerase with 32, but a switch to 38 takes place in the late exponential and early stationary phases. Regardless of the growth phase, expression from Pm remains dependent on 3MB3-activated XylS, and the transcription initiation point is unchanged (14, 15).
The earliest evidence of the involvement of 32 in Pm transcription came from the observation that, in an rpoH background, no expression of Pm took place in the exponential phase after induction, whereas expression increased during the stationary phase. Dependence on 38 was supported by the reduced transcription from Pm in the stationary phase in an rpoS mutant (14). In an rpoH-rpoS double mutant, only basal activity from the Pm promoter was detected along the growth curve. Analyses of transcription using combinations of mutant Pm promoters and mutant XylS proteins confirmed that the alternative factors interacted directly with the Pm promoter (14, 16). The increase in 32 activity required for transcription in the exponential phase was provided through induction of the heat-shock response by the presence of the effector 3MB, which is also required for activation of the positive regulator XylS (17). Microarray experiments with 3MB-induced P. putida (pWW0) cells further confirmed the fast (15 min) heat-shock response to this effector.4 From these findings, it follows that the Pm promoter should be recognized by two different RNA polymerases, and therefore it should accommodate the essential elements for recognition by both 38 and 32 in the same sequence stretch, although they are recognized at different moments during growth.
It is well established that factors play an essential role in programming gene expression, where the alternative subunits direct transcription toward specific gene sets according to growth and environmental conditions. The set of factors that forms the 70 family shares regions of extensive sequence homology and is organized in similar domains and subdomains. However, the DNA sequence recognized by the holoenzyme bearing each of the subunits is different. A number of 38-dependent promoters have been analyzed in an attempt to derive a consensus sequence. No clear -35 sequence has been defined, but the -10 region exhibits distinctive features (i.e. a C at position -7 in the -10 hexamer and CT at positions -13/-12). The region downstream from the -10 sequence in 38-recognized promoters is rich in adenines and thymines (18-21). On the other hand, a consensus sequence for 32-dependent promoters has been defined from the compilation of 18 Escherichia coli heat-shock promoters (22) (Fig. 1). As far as we know, none of these eighteen promoters required specific transcriptional activators, but heat-shock stabilization of 32 was the only requirement for activation in E. coli.
In this study, we analyzed the Pm promoter sequence to decipher the critical features that allow its unique response to RNA polymerases with either 32 or 38. We first precisely defined the -10 hexamer from mutant phenotypes, and we showed that, depending on the factor used by the RNA polymerase and despite the sole transcription start point, there are two different overlapping promoters at Pm. Our analyses made it possible to identify promoter bases that are critical for either one or both factors. We also analyzed the role and influence of relative factor abundance on Pm activity. A model on the functioning of Pm is proposed based on the analysis of the mutant promoters determined in this study.
|
EXPERIMENTAL PROCEDURES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
|
Construction of Mutant Pm Promoters—The Pm mutant promoters were generated by overlap extension polymerase chain reaction mutagenesis as described previously (25). The internal oligonucleotide primers used for mutagenesis exhibited one mismatch with respect to the wild-type sequence. The external oligonucleotides were the so-called M13 reverse primer (5'-CAGGAAACAGCTATGACCATG-3') or a universal primer (5'-GTTGTAAAACGACGGCCAGTG-3'). The template for each mutagenesis was 200 ng of pJLR100, and amplification conditions were as described by Higushi (26). After DNA amplification, the resulting DNA was digested with EcoRI and HindIII, and the 401-bp EcoRI-HindIII fragments containing mutant Pm sequences were inserted between the EcoRI-HindIII sites of pMD1405 to yield plasmids pMD::Pmx-zy, which carry in-frame Pm*::lacZ fusions, where x is the original base in Pm, located at position -z with respect to the transcription start site and mutated to y. All of the mutant Pm promoter sequences generated in this study were confirmed by DNA sequencing.
RNA Extraction and Analysis—Culture samples of 15 ml were harvested by centrifugation in disposable plastic tubes pre-cooled in liquid N2 and were kept at -80 °C until use. RNA was extracted using the TRI Reagent® from Molecular Research Center (Madrid, Spain) and modified as follows. The lysis step was carried out at 60 °C, and a final digestion step with RNase-free DNase was added at the end of the process. The RNA concentration was determined spectrophotometrically at 260 nm. Hybridization of the single-stranded 32P end-labeled DNA primer XylX (5'-GGGTCGGTGAACATCTCGCGCTTGC-3') (105 cpm/assay) complementary to the mRNA transcript originated from Pm and primer extension with avian myoblastasis virus reverse transcriptase were carried out as described previously (27). In all assays, 10 µg of total RNA was used as a template. cDNA products were analyzed in a urea-polyacrylamide sequencing gel, and gels were exposed to a phosphor screen (Fuji Photo Film Co, Ltd.) for 24-48 h. Phosphor screens were scanned using a phosphorimaging instrument (Molecular Imager FX, Bio-Rad). Data were quantified with Quantity One software (Bio-Rad).
-Galactosidase Assays—E. coli strains bearing the wild-type Pm::lacZ or mutant Pm*::lacZ fusions in pMD1405 plus a compatible plasmid bearing the wild-type XylS (pERD103, TABLE ONE) were grown overnight on LB medium containing the appropriate antibiotics. Three independent clones of each strain were used. Duplicate cultures were prepared by diluting cells from overnight cultures to 1/100. After 1 h at 30 °C, one of the duplicates was supplemented with 1 mM 3MB, while the other one was kept as an uninduced control. Samples for -galactosidase activity assays in the exponential phase were taken 30 min after induction when cultures had reached an OD660 = 0.1-0.3. Five hours after induction, the cultures had reached the stationary phase (OD660 = 2-2.5), and samples for -galactosidase activity were taken as above. -Galactosidase activity was determined in permeabilized whole cells according to Miller (28).
|
-galactosidase assays (see previous paragraph), and samples were taken in exponential and stationary phases. Pellets were resuspended in 1 ml of distilled water, cells were lysed by sonication, and the crude lysate was spun for 2 min at 12,000 x g. Supernatant aliquots of 100 µl were heated at 99 °C for 10 min to inactivate DNases. Four serial dilutions were prepared for each crude DNA sample. Real-time PCR was performed on an iCycler iQ detection system instrument according to the manufacturer's instructions. The PCR reaction (20 µl) was set up using the following reagents: 10 µl of 2x SYBR Green Supermix (Bio-Rad), 0.2 µl each of the forward and reverse primers for each gene (100 µM), 7.6 µl of water, and 2 µl of diluted DNA sample. The thermal cycling conditions used were: 3 min at 95 °C, followed by 35 cycles of 95 °C for 30 s, 60 °C for 20 s, and 72 °C for 30 s. A final melt curve was carried out to check the specific amplification of both genes. All reactions were run in triplicate. For each strain and condition, the two genes were analyzed.
Western Blots—Overnight cultures of E. coli strains were grown on LB medium containing appropriate antibiotics. Cultures were diluted 100-fold in the same medium, grown for 1 h at 30°C and then supplemented or not with 1 mM 3MB. After 30 min of incubation at 30 °C, the cells (100 ml) were harvested by centrifugation, and the pellet was resuspended and washed once in 1x M9 buffer (48 mM sodium phosphate, 22 mM potassium phosphate, 19 mM ammonium chloride, and 8.5 mM sodium chloride, pH 7). The resulting paste was stored at -20 °C until use. The same procedure was followed in parallel cultures after 5 h of induction, except that 10-ml samples were harvested. The cell pellets were resuspended in 2 ml of lysis buffer containing Tris (50 mM) pH 7.5, 50 mM NaCl, 2 mM EDTA, 4 mM -mercaptoethanol, and 1x CompleteTM protease inhibitor mixture (Roche Applied Science). The cells were lysed by sonication, and the insoluble fraction was discarded by centrifugation at 18 000 x g for 20 min. Care was taken to process all samples in parallel to avoid differences in protein dissociation. Aliquot fractions were analyzed on SDS-PAGE (12.5%) and transferred to a nitrocellulose membrane. To determine the relative fraction of each factor bound to the RNA polymerase core, the same samples were run in native 7% (w/v) PAGE and transferred as described above. The membranes were blocked for 3 h at room temperature with 5% nonfat dry milk in phosphate-buffered saline. Blots were incubated at 4 °C overnight with monoclonal antibodies against each E. coli subunit (purchased from Neoclone). Antibodies were diluted 1:3000 in the case of 70, 38, and antibodies and 1:1000 for 32 antibodies. Blots were washed with phosphate-buffered saline solution and incubated with goat anti-mouse immunoglobulin G + L conjugated with horseradish peroxidase (1:1000 dilution) for 1 h (Caltag Laboratories). The blots were developed with the SuperSignal® West Dura-extended duration substrate (Pierce). Chemiluminescent blots were exposed to auto-radiographic film for 30 s to 2 min. Time course experiments were carried out as described above, except that samples for extract preparation were taken 10, 30 and 60 min and 5 h after 3MB was added.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
Factors along the Growth Curve— factors bind to core RNA polymerase and allow the recognition of short nucleotide motifs located 10 and 35 bp upstream from the transcription start point of the promoters. Depending on the factor and the promoter, the so-called -10 box can be located at different positions between -5 and -14. In the Pm promoter, these sequences diverged considerably from consensus ones (Fig. 1). In this promoter, high expression levels were maintained throughout the growth curve by a switch from E32 in the exponential phase to E38 in the stationary phase. As a result, no significant decrease in activity was observed during growth, reflecting optimized coupling of the two RNA polymerases. The limits defining Pm transcription by each RNA polymerase can only be defined in mutant backgrounds. In our analysis, we used E. coli MC4100 as the wild-type strain and its isogenic derivatives RH90 (rpoS mutant) (30) and KY1429 (rpoH mutant) (31).
It is known that the promoter sequence is not the only factor that determines in vivo promoter selectivity by different RNA polymerases. Parameters, such as factor levels and availability or factor competition for the polymerase core, also play important roles (32). To pinpoint the conditions when the switch between factors occurred in the Pm promoter, it became necessary to determine their relative abundance in the wild-type and -deficient backgrounds under different growth conditions. To track the presence of factors 38, 32, and 70 in cells growing on LB medium in the presence or absence of 3MB, samples for Western blot analysis were taken in the early exponential or stationary phase as described under "Experimental Procedures." Similar levels of 70 were found in all three strains in both the exponential and stationary phases. These levels were independent of the presence or absence of 3MB (not shown).
In the exponential phase, 38 was only detectable in KY1429 (Fig. 2A). In the stationary phase, 38 was detectable in the parental strain MC4100 and in the rpoH mutant KY1429, whereas it was absent in the rpoS mutant RH90 (Fig. 2B). Interestingly, 38 stability was low in KY1429, as denoted by the multiple degradation bands that appeared in Western blots, particularly in the stationary phase. In this strain, in the presence of 3MB, 38 concentration greatly increased both in the exponential and stationary phases (Fig. 2, A and B). This could be due to impairment of the heat-shock response in the rpoH strain and, hence, to reduced induction of heat-shock proteins, such as ClpXP and DnaK (33, 34), which are responsible for the degradation and stability, respectively, of 38. To detect 32, doubled amounts of total protein were loaded into the gels. In the wild-type and rpoS mutant, 32 was detectable only in the exponential phase, regardless of the presence of 3MB, and was absent in the knock-out mutant KY1429 (Fig. 2C).
|
38 in response to the addition of 3MB was followed along a growth curve (Fig. 2D). In the wild-type strain, where 38 was absent in the exponential phase, the factor appeared 1 h after the addition of 3MB. In KY1429, 38 was detected as promptly as 10 min after induction. Heat-shock or stress induction of 38 in the exponential phase has been reported, and unlike 32, this response is not transient (33).
In an attempt to estimate changes in the fraction of each factor bound to the RNA polymerase core under our experimental conditions, samples from all three strains were taken in the exponential and stationary phases from cells grown in the presence and absence of 3MB, electrophoresed under nondenaturing (native) conditions, and blotted onto nitrocellulose membranes in quadruplicate. The membranes were then incubated with antibodies against the , 70, 38,or 32 subunits of RNA polymerase (Fig. 3). Free subunits and RNA polymerase holoenzymes could be identified when membranes incubated with antibodies against and 70 subunits were compared. Fig. 3, A and B, shows a slow migrating band that could be detected with antibodies against 70 and and was thus identified as the holoenzyme. It is interesting to note that, in the exponential phase, a portion of the subunit was found as a free subunit, whereas in the stationary phase, most of it appeared bound to the core. In contrast, 70 factor was mainly found as part of the holoenzyme.
Incubation with antibodies against 38 revealed two bands, one corresponding to free 38 and a second one migrating slowly (as described for the holoenzyme band) and thus corresponding to core-bound 38. The data shown in Fig. 3C confirmed that 38 was most abundant in KY1429 as a free subunit and also as part of the holoenzyme. In this strain, the addition of 3MB increased the fraction of 38 bound to the core. In the stationary phase, this factor was mostly found as a free subunit based on our analysis of Western blots. Our experimental conditions did not allow the detection of 32 in the native electrophoresis.
|
32 and 38 in MC4100 in the exponential and stationary phases. In the rpoH mutant KY1429, 38 appears to become available earlier during growth, both free and as part of the holoenzyme. As a consequence, E38 would be able to transcribe Pm earlier during the growth period than in the wild-type strain MC4100 (see "Expression from Pm in an 32-deficient Background").
Scanning Mutagenesis of the RNA Polymerase Binding Region at Pm— To identify the critical nucleotides able to confer specific recognition of Pm by either 32 or 38, we generated all possible single point mutations between positions -7 and -18. The region under analysis was extended to -18, because the Pm sequence resembles the 32 consensus sequence between positions -14 and -17 (CCCC) (35) (Fig. 1). The mutant promoters were fused to lacZ in plasmid pMD1405, and the level of expression in the early exponential phase of growth (culture turbidity at 660 nm between 0.1 and 0.3) and the stationary phase (culture turbidity at 660 nm of about 2-2.5) was determined in the wild-type strain MC4100 (xylS gene in pERD103) and in the presence of the effector molecule 3MB. In this strain, 3MB-dependent -galactosidase levels for the wild-type Pm promoter were 1,300 MU in the early exponential phase and 15,000 MU in the stationary phase (Fig. 4). Fig. 4A shows -galactosidase activity levels obtained with each single mutant promoter in the exponential phase. Pm nucleotide positions in the mutated RNA polymerase binding region could be grouped according to the expression pattern of the corresponding mutant: 1) positions -7, -11, and -12, where any change was critical regardless of the specific base that was introduced and completely abolished Pm activity; 2) positions -8 to -10, where most mutations had a severe effect upon activity; and 3) positions between -13 and -18, where mutations had little effect on Pm activity in this growth phase.
The situation was different when Pm-dependent -galactosidase activity was determined in the stationary phase (Fig. 4B). We found low levels of transcription with the Pm mutant promoters that had point mutations at -7, -11, and -12, and a modest level of activity was detected only in the -12T 
C Pm mutant. On the other hand, activity recorded with mutation -8C A was almost double the level found in the wild type. One of the possible changes at positions -9 and -16 led to a 65% decrease in activity, whereas all mutations at positions -17 and -18 produced decreases in activity between 25 and 50%. The remaining mutants behaved like the wild type, showing a higher level of expression than in the early exponential phase. It is worth noting that the -galactosidase levels obtained in the absence of the effector 3MB were negligible.
These results were interpreted as evidence that the RNA polymerases acting in the exponential and stationary phases recognize the same nucleotides at -7, -11, and -12 but recognize additional positions differently. In particular, the low level of activity with certain mutants at -8 and -9 in the early exponential phase, when 32 was used to transcribe Pm, indicates that these two nucleotides are critical for the 32 factor. The -8C A change rendered a mutant promoter that was fully active in the exponential phase and exhibited a 2-fold higher expression level than the wild-type promoter in the stationary phase (Fig. 4). This increased transcription from this mutant promoter in the stationary phase was initially attributed to E70, because the mutation improved the -10 region to those recognized by E70. To confirm this possibility, all Pm mutants were analyzed in the rpoS-rpoH double mutant SM25 (15), where xylS was provided in the gentamicin-resistant vector pDCXylS. All Pm mutants showed activity levels below wild-type Pm in the absence of both subunits (not shown), thus suggesting that 70 was not responsible for these transcription levels. We therefore concluded that the sequence in the -8C A mutant was significantly improved for 32 and especially for 38. The results shown in Fig. 4 suggest that, although wild-type G at position -9 is important for recognition by 32, a change to A enhances 38 recognition of the promoter (Fig. 4B). Overall, changes in the promoter sequence toward optimization of recognition by 38 seem to be unfavorable for E32. The actual wild-type sequence may therefore be a compromise between the ability to respond to 38 in the stationary phase and the requirement to respond to 32 in the early exponential phase.
Sequence Determinants of Pm Expression in the Exponential Phase— E. coli rpoS strain RH90 was transformed with plasmid pERD103 (xylS) and pJLR107 (Pm:lacZ) or its derivatives containing mutant promoters, and we determined -galactosidase activity in the early exponential growth phase as indicated under "Experimental Procedures." We had previously shown that activity from the wild-type Pm promoter in the exponential phase was almost unaffected in a mutant lacking 38 (14), because under these conditions, Pm activity relies mainly on 32, which expression is maintained (Fig. 2C). Wild-type Pm activity in RH90 in the exponential phase was 1500 MU (115% in Fig. 5A). When we examined activity from the mutant promoters in the early exponential phase when only 32 was available for transcription from Pm, very low activity was found for all of the mutant promoters with mutations at positions -7, -9, -11, -12, -15, and -17. When mutations were introduced at position -8, -10, -13, -14, -16, and -18, -galactosidase activity varied depending on the specific mutation that was introduced. These results indicate that not only the -10 hexamer but also the bases upstream from the -10 motif (namely Cs at positions -14 to -17) are crucial for 32 recognition. The presence of an A at position -8 allowed the levels of expression to approach those in the wild type, thus confirming the results obtained with strain MC4100. In contrast, G at position -13 or -14 produced -galactosidase levels of up to 50% of those in the wild-type Pm. Interestingly, two mutations at position -10 led to wild-type activity levels, indicating that the changes were irrelevant for 32 recognition in Pm. However, a change to C prevented recognition by 32.
|
|
|
32 at Pm as 5'-(H)CCC(S)(K)TA(D)G(M)T-3' (positions -18 to -7), where T at positions -7 and -12, G at position -9, A at position -11, and C at positions -14 to -17 are pivotal for E32 transcription. Although similar to the functional 32 binding sequence deduced from the E. coli groE promoter (36), this sequence shows a number of differences, particularly and most importantly the identification of position -7 as relevant for recognition by 32.
Wild-type Pm activity in the rpoS mutant background was 60% of its value in MC4100 (7,000 MU) in the stationary phase. In general, in this mutant background, most of the mutants tested exhibited low activity levels. As mentioned above, expression in the -8C A mutant remained at the wild-type level (Fig. 5), and mutations -10G T or -10G A led to promoters with higher levels of expression than wild type.
Expression from Pm in an 32-deficient Background—To identify the promoter positions that were important for E38 recognition, we decided to follow Pm expression in a mutant lacking the E32 polymerase involved in expression in the exponential phase. In this 32-deficient background, most of the transcriptional activity measured at Pm is presumably due to E38, which is already present in stationary phase (Fig. 2). Pm expression in the MC4100 rpoH derivative KY1429 was very low for the wild-type promoter in the exponential phase (75 MU) and increased in the stationary phase (2700 MU) (Fig. 2) to levels below those seen in wild-type MC4100. When we measured mutant Pm expression in this strain, we observed that mutations at positions -7, -11, and -12 in Pm resulted in similar basal levels of expression to those observed in the wild-type MC4100 and RH90 strains (Fig. 6). In addition, the mutations at position -10 in Pm, which behaved like the wild type in RH90, were expressed at a lower level than those in the wild-type strain, particularly in the stationary phase. This was also the case for mutant -10G C. These results suggest that position -10 is important for 38 recognition. However, when mutated, most of the remaining positions rendered a promoter with increased expression in the exponential phase. This was the case for all mutations at position -9, the mutation at -8G, and at least one of the mutations at positions -14 to -17 (Fig. 6A). These differences were maintained in the stationary phase, when all mutations except those at positions -7, -11, and -12 led to levels of expression equal to or higher than those induced by the wild-type Pm. This general increase in activity with respect to the wild type both in the exponential phase, when 38 was already present in this strain, and in the stationary phase, when the highest levels of 38 were seen in KY1429 (Fig. 2), indicates that the wild-type Pm promoter was far from optimal for recognition by E38 and that the absence of 32 competition for both the RNA polymerase core and DNA binding would allow transcription by E38 to increase. The previously mentioned lack of transcription in a double mutant rpoS-rpoH rules that these levels were because of residual 70 activity.
|
-galactosidase assays we used were not able to detect. To verify that the results shown in Figs. 4, 5, 6 were actually derived from Pm activity and not from any accidental promoter created as a consequence of the mutagenesis, the transcription start site of the most relevant mutant promoters in our analyses was determined in the wild-type and mutant strains. Fig. 7 shows that, in all mutants tested, the transcription start site was the canonical Pm start site, and activity levels correlated well with -galactosidase activity. Transcription always required activation by 3MB-activated XylS, which probably guarantees the proper positioning of RNA polymerase for transcription activation. |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
32 and E38, involved in Pm expression.
A number of 38-dependent promoters have been reported to be also transcribed in vivo by a second RNA polymerase depending on physiological conditions. However, in all cases described so far, E70 is responsible for this alternative transcription. In the csgBA promoter, H-NS influences the use of either E38 or E70 to transcribe this promoter (37), whereas in the dps promoter E70 directs transcription in the exponential phase and requires OxyR. In the stationary phase, however, transcription from the dps promoter is 38-dependent and OxyR-independent (38). In contrast with these promoters, which use alternative regulators depending on the growth phase, Pm always requires XylS activated by 3MB to mediate transcription with either E32 or E38 regardless of the growth phase.
Our results with the Pm promoter in the wild-type and mutant strains show that E32 is the major polymerase involved in transcription of Pm. In the rpoH mutant KY1429, expression levels were only 7% of those in the wild type. Furthermore, in KY1429, the levels of expression from Pm in the stationary phase were relatively low, which suggests that the E38 polymerase transcribed poorly the Pm promoter (Fig. 6). In fact, expression was low despite the presence of E38 in the early phase of growth (Fig. 2).
Defining the Downstream Sequences Recognized by 32 and 38—To define the downstream sequences recognized by 32 and 38, we used a series of Pm mutant promoters and tested expression in different genetic backgrounds deficient in either of the alternative factors. The rationale behind this approach was the following. If Pm is transcribed with E32 in the exponential phase and E38 in the stationary phase, one would expect the nucleotides at Pm that are important for transcription by E32 to be detected in the exponential phase as inactive mutants in MC4100 (wild-type background). In contrast, nucleotides involved in Pm recognition by E38 would be inactive in the stationary phase. Along these same lines, mutations at the Pm promoter involved in 32 recognition should also be inactive in RH90 (32-proficient, 38-deficient) in the exponential phase when E32 plays an important role in Pm transcription. Moreover, mutations in Pm at positions relevant for E38 recognition should be affected in KY1429 (38-proficient, 32-deficient) in the stationary phase. In addition to the above considerations, in our analysis, we took into account the levels of 38 and 32 in the different genetic backgrounds used in this study that would influence the availability of 32 and 38 for binding to the core RNA polymerase (Fig. 3). No less important is the fact that mutations in the -10 element could generate recognition sites for 70. These issues were taken into consideration in our analysis of each mutant.
The precise definition of the -10 hexamer in promoters transcribed by RNA polymerase with the 70 family of alternative factors is elusive, because the distance to the transcription start site can vary from 4 to 8 bp and also because this region is usually rich in adenines and thymines; so different hexamers can be proposed (18-21, 39, 40). In the case of Pm, because of the profusion of adenines and thymines between positions -5 and -13 (Fig. 1), several putative -10 hexamers can be hypothesized. In our analysis, any mutation at positions -7T, -11A, and -12T dramatically decreased Pm activity in all the genetic backgrounds we tested. This allowed us to precisely define the -10 recognition element of Pm as TAGGCT (positions -12 to -7). The presence of T in the first and last position of the hexamer is in agreement with observations reported for other promoters transcribed by 38 (35, 41, 42).
Mutations within the -10 Element—The consensus sequence shown for 32 in Fig. 1 was first defined some twenty years ago. Recently, scanning mutagenesis of the E. coli groE promoter confirmed the earlier consensus sequence of the 32 -10 element (36) (Fig. 1). The alignment of the Pm promoter sequence with the 32 consensus sequence supports the recognition of Pm by E32. Two mutations at positions -8 (-8C G and -8C T) and one at -9 (-9G A) showed the expected phenotype for a relevant 32 binding position, low activity in the exponential phase and unaffected in the stationary phase in the wild-type MC4100. These Pm mutants were also inactive in RH90 in the exponential phase. However, the behavior of two of these mutants in the 32-deficient KY1429 background merited further investigation. The -8C G and -9G A mutants showed high activity levels in this strain, suggesting that E38 was able to transcribe from this promoter in the absence of 32. Published reports of the changing levels of 38 throughout the growth phase or at different growth temperatures are discrepant and seem to vary from strain to strain (43, 44). Our results in Fig. 2A showed 38 to be present at significant levels in the early growth stages both in MC4100 and KY1429, although in the latter, 38 was always present and its levels were higher and detectable even in the absence of 3MB (Figs. 2 and 3). The 38-dependent transcription of Pm mutants -8C G and -9G A was not observed in MC4100, presumably because the presence of 32 decreased RNA polymerase core availability for 38.
Specific features common to 38-dependent promoters have been established from compilations of an increasing number of these promoters (18, 21) from mutagenic analyses both in vivo and in vitro and from the analysis of artificial promoters containing putative 38-dependent sequences (45, 46). These promoters usually lack a defined -35 element, exhibit four highly conserved nucleotides in the -10 box (21) (Fig. 1, underlined), and generally have a TC at positions -14/-13 (47). The Pm promoter shares five of eight positions with the recently updated 38 consensus sequence (21) (Fig. 1). However, the promoter lacks the main features described by Lacour et al. (20) for promoters that are exclusively recognized by E38, i.e. a C in the first position in the -10 hexamer.
In the stationary phase, the set of mutations in -8 and -9 allowed activity levels that approached the wild-type levels in MC4100, and that surpassed wild-type levels in KY1429, suggesting that almost any change in Pm would favor E38 recognition. However this is not the case with mutations -9G C or -8C T, which do not improve E38 recognition. In RH90, the levels were not maximal, as a consequence of the negligible activity in the exponential phase, which curtailed -galactosidase accumulation. Mutations in the -10 position produced a similar expression pattern in MC4100 but exhibited negligible activity in the KY1429 background, although two of the mutants yielded promoters that were fully active in RH90. These findings suggest that this position was not very important for 32 but was necessary for 38 recognition. In fact, this was the only position where mutants showed a 38-dependent phenotype. This is striking, because guanosine was the least common nucleotide in the -10 hexamer in a recent compilation of 38-dependent promoters, although the substitution of a canonical consensus nucleotide for a G at this position in an artificial 38-dependent promoter had no effect on 38 DNA binding (21).
We showed previously that, in a double rpoS,rpoH mutant, transcription from Pm was almost negligible, which was taken as an indication that, in the wild-type Pm promoter, E70 played a minor role (15). Mutations at -7 and -11 yielded a nonfunctional Pm promoter in all of the backgrounds tested in this study. This was also true for position -12, except that when the -12T position in Pm was mutated to a C, the resulting promoter retained 25% of the activity in the stationary phase in all backgrounds (Figs. 4B and 5B). Position -12 is known to be crucial for RNA polymerase recognition. The -12T C mutation gives rise to a mutant Pm promoter with a low activity, which however, was still recognized to a certain extent by both polymerases. We also found that the -8C A change resulted in a high level of Pm expression in the stationary phase in the wild-type MC4100 (Fig. 3B) and in the 38-deficient mutant. As for mutant -12T C, this mutant is transcribed by both polymerases but with a high efficiency, especially by E38. In this case, the Pm sequence was improved for both holoenzymes.
Taken together, these results show that, in addition to the pivotal positions -7, -11, and -12, which are crucial for the transcription of Pm with any RNA polymerase, positions -8 and -9 are critical for 32 recognition but are much less important for 38 recognition, whereas the -10 position is key for 38 recognition.
Mutations Upstream from the -10 Element—When we applied the same rationale as above to mutants at positions -12 to -18, we observed that no mutation had a significant effect on activity in the exponential phase in MC4100. However, in RH90, which lacks 38, all Pm mutant promoters bearing mutations between positions -13 and -17 showed between 10 and 50% activity of the wild-type Pm in the same background. Moreover, the Cs present at positions -14 to -17 could be aligned with the 32 consensus (Fig. 1). The high activity levels of these mutants in MC4100 suggest that, in this wild-type strain, E38 takes over transcription from these promoters. This hypothesis is further supported by the considerable levels of -galactosidase obtained in KY1429 in the exponential phase.
These results suggest that mutations between -13 and -17 resulted in mutant promoters better recognized by E38 in both the exponential phase and stationary phase (e.g. all changes at positions -17, -15A, -15G, -14T, -13C). However, this was only observed in the absence of 32 (strain KY1429) and was confirmed by the lower activities obtained from these mutant promoters in RH90 in the stationary phase. In MC4100, the activity of these promoters is similar to wild-type Pm, probably because 38 must compete with 32 for binding to the RNA polymerase core.
We showed that Pm is transcribed by two RNA polymerases, E38 and E32, from the same transcriptional start point in a process that is strictly dependent on XylS activated by 3MB. Our results suggest that deviations from the 38 consensus are required to adapt Pm to the demands of its function, i.e. recognition by E32 as a rapid response to the presence of 3MB. This response is strong but transient, so the promoter must be optimized for recognition by this polymerase. In fact, the Pm sequence is closer to the 32 consensus, so changes in this sequence are not tolerated. As a result, Pm is adjusted to ensure almost optimal recognition by this less abundant polymerase. After this initial reaction of the cells to the presence of the aromatic effector, activity with E38 ensues and is subsequently maintained. This polymerase does not rely so strictly on the promoter sequence and is able to promote transcription from the modestly conserved 38 binding sequence present in Pm.
The Pm promoter always requires the presence of activated XylS to be transcribed. It has been shown that the nature of the regulator that activates Pm strongly influences the factor used in transcription. Wild-type XylS seems to be designed to prevent E70 accessibility to Pm (16). This implies an additional benefit for Pm transcription control, namely XylS would restrict transcription to situations in which the heat shock response and, hence, E32 abundance are triggered by the presence of aromatic pathway substrates.
|
FOOTNOTES |
|---|
1 Recipient of an I3P contract from the European Social Funds and a fellowship from the Junta de Andalucía (Andalusian Regional Government, Spain).
2 To whom correspondence should be addressed: Estación Experimental del Zaidín, CSIC, C/o Profesor Albareda 1, E-18008 Granada, Spain. Tel.: 34-958-181600; Fax 34-958-129600; E-mail: silvia{at}eez.csic.es.
3 The abbreviations used are: 3MB, 3-methylbenzoate; MU, Miller units.
4 P. Domínguez-Cuevas, J. E. González-Pastor, S. Marqués, J. L. Ramos, and V. deLorenzo, submitted for publication.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
20% of the activity in MC4100). Vector pMD1405 copy number, as determined by real-time PCR (see "Experimental Procedures"), both in exponential and stationary phases was the same in MC4100 and KY1429.