Constitutive Function of a Positively Regulated Promoter Reveals New Sequences Essential for Activity

A consensus “-10” recognition sequence for RNA polymerase was created at the positively regulated X Pre promoter by introducing three single base pair mutations. This altered promoter, Pre:, functions constitutively in vivo and in vitro at high efficiency despite very poor consensus “-35” region sequence homology. We examined the influence of the -35 region sequence information on promoter function by shifting the wild type -35 region 2 2 base pairs relative to the -10 region consensus sequence and by completely replacing it with alternative DNA sequences. In every case, the altered Pre* promoters retained transcriptional activity although differences in their transcriptional efficiencies were observed. Apparently the Pre* promoter does not require specific -35 region sequences for constitutive promoter activity, although the -35 region sequences can modulate overall promoter strength. In addition, by point mutation analysis we have identified bases immediately upstream of the 10 hexamer which are essential for constitutive function of the Pre* promoter. We propose that these mutants define an extended -10 region at Pre* that compensates for its poor -35 region sequence information by providing critical contacts that stabilize productive RNA polymerase binding.

A consensus "-10" recognition sequence for RNA polymerase was created at the positively regulated X Pre promoter by introducing three single base pair mutations. This altered promoter, Pre:, functions constitutively in vivo and in vitro at high efficiency despite very poor consensus "-35" region sequence homology. We examined the influence of the -35 region sequence information on promoter function by shifting the wild type -35 region 2 2 base pairs relative to the -10 region consensus sequence and by completely replacing it with alternative DNA sequences. In every case, the altered Pre* promoters retained transcriptional activity although differences in their transcriptional efficiencies were observed. Apparently the Pre* promoter does not require specific -35 region sequences for constitutive promoter activity, although the -35 region sequences can modulate overall promoter strength. In addition, by point mutation analysis we have identified bases immediately upstream of the -10 hexamer which are essential for constitutive function of the Pre* promoter. We propose that these mutants define an extended -10 region at Pre* that compensates for its poor -35 region sequence information by providing critical contacts that stabilize productive RNA polymerase binding.
The positively regulated X Pre promoter is not recognized directly by RNA polymerase. Promoter function is completely dependent on the phage transcriptional activator protein, cII (1,2). Activation of Pre by cII plays a vital role in achieving the delicate balance between lytic and lysogenic development observed for phage X (3)(4)(5).
Analysis of the DNA sequences of promoters recognized by Escherichia coli RNA polymerase have identified two conserved 6-base pair sequences located approximately 10 and 35 bases upstream from the transcription start site. They are referred to as the "-10" and "-35" promoter regions, respectively (for reviews see Refs. [6][7][8]. The importance of these hexanucleotide regions has been demonstrated by extensive mutational studies. In fact, almost all of the promoter point mutations that have been characterized map to the -10 or -35 hexanucleotide regions (8,9). Another conserved promoter feature is the spacing distance between the -10 and -35 sequences. This spacer region has a conserved length of 17 base pairs and demonstrates little, if any, DNA sequence specificity (10)(11)(12)(13)(14)(15)(16)(17).
* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $This work was part of S. Keilty's thesis project at the State University of New York at Albany.
At the level of DNA sequence, the X Pre promoter has poor homology with consensus promoter sequences (18,191. In the -10 region, Pre has only three of the six conserved bases, and in the -35 region, 17 base pairs upstream, there is essentially no homology. Thus, it is not surprising that the X Pre promoter is not recognized directly by RNA polymerase and that transcription from Pre is dependent upon the X activator protein, cII. The cII protein specifically recognizes a tetranucleotide repeat sequence, TTGC(NG)TTGC, separated by 6 base pairs, (N6), which corresponds to the -35 region of the X Pre promoter (2,20). The interaction of cII protein with this sequence is both necessary and sufficient for productive RNA polymerase binding and efficient transcription from the Pre promoter.
More than 35 point mutations in the X Pre region have been characterized (21). However, despite considerable effort no cII-independent promoter mutations have been obtained. Presumably this is because a single point mutation would not be sufficient to allow independent RNA polymerase recognition. This contention is supported by the lack of consensus information at the Pre promoter site and the fact that a single base change toward consensus in either the -10 or -35 promoter region would not constitute a sufficient increase toward consensus. A truly constitutive X Pre promoter mutant would probably require multiple point mutations within the confined regions of the -10 and/or -35 hexamer sequences.
In an effort to construct an efficient constitutive derivative of the X Pre promoter we increased its homology with consensus sequences in the -10 region. Three point mutations were introduced into this region to create a perfect consensus -10 region sequence (TATAAT). We demonstrate that this alteredpromoter, R e * , functions independently of the activator protein a t high efficiency both in vivo and in vitro. This efficient constitutive promoter activity of Pre* occurs despite the fact that this promoter has essentially no homology with consensus sequence information in the -35 region. To determine the influence of -35 region sequences on the constitutive function of the Pre* promoter we replaced the original X -35 region sequence with several alternative sequences. We show that no specific bases are required in any of the six positions of the -35 region to promote constitutive transcription from Pre*. Moreover, we identify bases just upstream of the -10 region which have become crucial for constitutive function of the Pre* promoter.
Plasmid Constructions-The galK expression vector used in this study, pKO-SK, is a modified version of the plasmid pKO-11 in which the Nde-1 and EcoRI restriction sites at positions 1936 and 4000, respectively, were destroyed by fill-in reactions with the Klenow fragment of DNA polymerase (23). The resulting vector, pKO-SK, is suitable for both in uiuo and in uitro transcriptional studies. The h on a 95-base pair Dde-1(38342)/Dde-1(38437) fragment which had Pre promoter region extending from position +2 to -93 was cloned been filled in with the Klenow fragment of DNA polymerase and ligated into the Sma-l site of the galK expression vector, pKO-SK, to construct the plasmid pKO-Pre (Fig. 1, A and B). The plasmid pKO-Pre was cleaved at the unique EcoRI and Nde-1 sites to remove the wild type promoter sequences between positions +2 and -16. Then to construct the plasmid pKO-Pre*, we replaced the wild type promoter sequence from positions +2 to -16 with a synthetic oligonucleotide that contained the identical promoter sequences except for three single base changes in the -10 region. The introduction of these point mutations at positions -12, -10, and -9 created a perfect consensus -10 promoter region at Pre*. Another synthetic oligomer with the wild type X Pre sequence from positions +2 to -16 was also reconstructed to serve as an exact control for the Pre* promoter construct. This plasmid is called pKO-Pre+.
To create the %base pair insertion and deletion mutants at the Pre and Pre. promoters, the position of the -35 region was shifted relative to the -10 region by linearizing each plasmid at the unique Nde-1 site in the intermittent region. The insertion mutants, pKO-Pre+2 and pKO-Pre*+2, were constructed by filling in the 2-base pair overhang with the Klenow fragment of DNA polymerase. The deletion mutants, pKO-Pre-2 and pKO-Pre+-2, were constructed by removing the 2-base pair overhang with Mung bean nuclease. These constructions were all confirmed by Maxam and Gilbert DNA sequence analysis (24) (data not shown).
Several derivatives of the Pres promoter with the alternative -35 region sequences were constructed. The first derivative was made by filling in the Nde-l and EcoRI ends of the plasmid pKO-Pre* with the Klenow fragment of DNA polymerase. Then this 18-base pair fragment which contained the promoter sequences from positions +2 to -16 and the consensus -10 sequence was ligated directly into the Sma-1 site of the vector pKO-SK to generate the plasmidpK0-Pre*l. Two other plasmids with alternative -35 regions, pKO-Pre*2 and pKO-Pre*3, were constructed by linearizing the Pre* plasmid with Nde-l and inserting a 555-base pair Nde-1(36115)/.Vde-1(36670) fragment from the phage h in both orientations. These constructions were verified by fine restriction mapping (data not shown).
Three derivatives of the Pre* promoter with single point mutations just upstream of the consensus -10 region were constructed using the same procedure as described for the construction of pKO-Pre*. The plasmids, pKO-P14T*, pKO-P14C*, and pKO-P15C*, contain the following base substitutions, G to T at position -14, G to C at J B

"-35"
Nde-I "-10" -GTTTGTTTGCACGAACCATATGTAAGTATTTCClTAG 3 " r FIG. 1. pKO-SK and the X Pre promoter region. A, this diagram of the plasmid pKO-SK shows the Pre promoter region insert (not to scale) with the direction of transcription indicated by an arrow, the E. coli galactokinase gene, gal K downstream of the promoter, several unique restriction sites, the ampicillin resistance gene, amp, and the origin of replication, ori. The size of the pKO-SK vector is 4002 base pairs. B, the base sequence of the noncoding strand of the X Pre promoter region insert extends from position -93 to +2 relative to the transcription start site which is denoted with an arrow. The -10 and -35 region hexamers are indicated as is the unique Nde-l restriction site in the spacer region. position -14, and T to C at position -15, respectively. These constructions were verified by fine restriction mapping and loss of the unique Nde-1 restriction site.
The plasmids pKO-Pre andpKO-Pre+2 were modified by inserting a 190-base pair DNA fragment that contained the transcription terminator, toop, into the unique Nru-l site 180 base pairs downstream of the transcription start sites of the promoters. The resulting plasmids, pKO-Pret and pKO-Pre+2t, respectively, were used as supercoiled DNA templates for in vitro run-off transcription experiments, described below.
Transcription in Vitro-The in uitro run-off transcription reactions were performed as described previously (1,2). All transcription reactions contained 1.0 pg of plasmid DNA linearized at the unique Nru-1 site (180 base pairs downstream of the promoter regions) unless otherwise noted. The transcription mixtures contained 0.3 mM ATP, CTP, and GTP and 0.2 mM UTP, 20 pCi of a-labeled [32P]UTP (specific activity, 410 Ci/mmol), 25 mM Tris-HC1, pH 7.5, 80 mM KCl, 10 mM MgCl,, 0.1 mM EDTA, and 1.0 mM dithiothreitol. Purified h cII protein was added to the reactions as indicated. Reactions were initiated by addition of limiting (0.5 unit) or excess (2.0 units) amounts of E. coli RNA polymerase purchased from Pharmacia P-L Biochemicals (specific activity, 342 units/mg) and incubated at 37 "C for 15 min. Reactions were stopped by addition of 50 p1 (8 M urea, 0.5 X TBE, 0.05% sodium dodecyl sulfate, 0.25% xylene cyanol/ bromphenol blue), denatured at 65 "C for 2 min, and then 5-10 pl was loaded onto a 5% polyacrylamide urea gel and subjected to elecrophoresis.
Galactokinase Assays-The pKO-SK plasmid derivatives were transformed into isogenic cII+ or cII-X lysogens of the galK deletion strain, uc6183. The lysogens had either a cII+ or cII3067 genotype, the temperature-sensitive repressor mutation, cI857, and additional mutations in both phage arms to prevent cell lysis upon induction. The cells were grown at 32 "C to an optical density of 0.5 and then thermally induced at 42 "C for 30 min (22). Using this system, the activity of the various promoter constructions was quantified in both the presence and absence of cII protein. Cell lysates were prepared by sonication and assayed after incubation at 30 "C for 15 min as described previously (23). Units are represented relative to maximally activated X Pre which produces an average of 565 galactokinase units and is equal to 100%.

RESULTS
Pre* Functions Constitutively in Vitro-In an effort to isolate a constitutive derivative of the X Pre promoter we introduced three base changes in the -10 region creating a consensus -10 sequence (TATAAT) (Fig. 2). This altered promoter, Pre*, was tested to determine if the increase in -10 region homology was sufficient to alleviate the cII dependence of the promoter, i.e. create a constitutive promoter. A wild type X Pre sequence was also reconstructed to serve as a control for these experiments (see "Experimental Procedures" for details).

T T G C G T T T G T~T G C A C G A A C C A~A T G T T A T A A T T T C C T T A G
p r r -2

P r r ' l C G T I G C G T T T G T T I G C A C G A A C~T T A T A A T T T C C T T A G
-Pre.1

G G C T G C A G G T C G A C G G A T C C C C T A T G T T A T A A T T T C C T A G -~r e . 2 I C T T G G G C C G A C A T T G T C A T C A T A T G T T A T A A T I T C C T T A G
pri..3

P l 4 T ' I T G C G T T T C T T T G C A C G A A C C A T A T T T T A T A A T T T C C T A G -~1 4 c ' T T G C G T T T G T I T G C A C G A A C C A I A T C I T A T A A T T T C C T T A G
P15C' These constructions were tested for transcriptional activity using linearized templates in standard in vitro run-off transcription assays. We found that the wild type -10 region construct, pKO-Pre+, retained complete cII dependence and functioned identically to the original X Pre construct, pKO-Pre. Therefore, it was an appropriate control for wild type Pre function. In contrast, the Pre: promoter functioned independently of cII protein, and we found that its activity was approximately equivalent to the activity of the X Pre promoter maximally activated by cII protein (Fig. 3). In transcription reactions using excess RNA polymerase , cII protein had no additional effect on transcription from the Pre* promoter (data not shown). However, in transcription reactions containing limiting amounts of RNA polymerase, the addition of cII protein reproducibly increased the level of transcription approximately 2-fold (Fig. 3). Apparently cII protein still recognizes the Pre: promoter and helps it compete for enzyme when RNA polymerase is limiting.

T T G C G T T T G T T T G C A C G A A C C A T A C G T T A T A A T T T C C T T A G
Pre* Functions Constitutively in Vivo-The same plasmid DNAs that were used as templates in the in vitro run-off transcription experiments were also used to examine and compare Pre and Pres function in vivo. These vectors are derivatives of the galK fusion vector system, and thus their relative promoter efficiencies can be compared directly in vivo by assaying galactokinase activity (25). Each plasmid was transformed into the appropriate galK-host to allow measurement of galK enzyme activity produced in the presence and absence of cII protein (see "Experimental Procedures" for details). The results obtained in vivo were completely consistent with those found in vitro. Whereas expression from 'the wild type Pre promoter was completely cII dependent, the Pre: promoter functioned independently of cII protein (Table  I). Moroever, the Pre* promoter expressed galactokinase a t essentially the same levels as did the cII-activated X Pre promoter. Clearly the three base changes in the -10 region

TABLE I Summary of the relative promoter activities awayed in vivo
The relative transcriptional efficiencies of the indicated promoters fused to the E. coli galactokinase gene on the pKO series multicopy plasmids were determined by assaying galactokinase expression in isogenic lysogens with either cII+ or cIIgenotypes as described in Refs. 22 and 23. Briefly, lysogens were grown to log phase a t 32 "C and then shifted to 42 "C to induce cII gene expression. Then 30 min after induction, cell extracts were prepared and galactokinase activities determined. The percent galactokinase expression is given relative to fully cII-activated X Pre which is 100% and equal to 565 units of galactokinase/l5 min at 30 "C. Each number represents the average of a t least three independent experiments using different cell extract preparations in which the calculated galactokinase units varied by less than 10%. make the Pre* promoter function constitutively and efficiently.
It is of interest to note that in vivo the activity of the Pre: promoter was increased almost 2-fold in the presence of cII protein ( Table I). This increase is comparable with that observed in vitro when the concentration of RNA polymerase in the transcription reactions was limiting. Apparently cII protein helps the Pre* promoter compete for limiting RNA polymerase in vivo as well as in vitro. Effects of Spacer Length on Promoter Function-It was somewhat surprising that the constitutive Pre* promoter functioned as efficiently as the fully cII-activated X Pre promoter. The cII-activated Pre promoter is known to be a rather efficient promoter, and the Pre: promoter functions as efficiently although it has essentially no homology with the -35 region consensus sequence. In order to examine the contribution of sequences in the -35 region to Pre: function we made alterations in the length of the spacer region separating the -10 and -35 promoter regions. It is well established that promoter activity is extremely sensitive to alterations in the spacer length (10)(11)(12)(13)(14)(15)(16)(17). At Pre* we altered this distance by insertion or deletion of 2 base pairs from the Nde-1 site within the spacer region to construct the pKO-Pre:-2 and pKO-Pre:+2 promoters, respectively (see "Materials and Methods" for details). These constructions were examined as above for their effect on Pre* function both in vivo and in vitro.
Transcription experiments carried out in vitro and in vivo (Fig. 4, Table I) indicated that both the Pre*+2 and Pres-2 promoters retained the ability to function constitutively and efficiently. In fact, the Pre:+2 promoter was more active than the Pre: promoter (Table I). This increased activity is probably due to the fact that the 2-base pair shift in the Pre:+2 promoter places the sequence (TTGTTT) in the appropriate -35 region position (Fig. 2) and thereby increases the homology with the consensus sequences (TTGACA). The Re*-2 promoter is almost as active as the Pre* promoter (Table I) base pair deletion (Fig. 2). These results clearly indicate that the Pre* promoter has a most unusual flexibility in its -35 region sequence requirements for obtaining relatively efficient promoter function.
We also examined the ability of these altered promoters to function in the presence of cII protein. In vivo the levels of transcription from both the Pre*-2 and Pre*+2 constructs were reproducibly reduced relative to the constitutive levels (Table I). However, in vitro the levels of transcription could be reduced by about 50% when saturating concentrations of cII protein (400 nM) were added to the transcription reactions (Fig. 4A). Apparently cII protein competes with RNA polymerase for binding a t these promoter sites and thus has a weak repressive effect. Presumably this is due to the 2-base pair shift which alters the juxtapositioning of the -35 region and flanking cII binding site relative to the -10 region and transcription start site. This result suggests that the correct spacing may be very important for cII-activated transcription.
T o examine the importance of correct spatial positioning for &-activated transcription we made analogous insertion and deletion mutations in the wild type X Pre promoter resulting in the Pre+2 and Pre-2 promoters, respectively (see "Materials and Methods" for details). Neither of these promoter signals responded to cII activation in vivo or in vitro confirming that the correct spatial positioning of the clI binding site relative to the -10 region and transcription start site is critical for cII activation (Fig. 4B, Table I).
Surprisingly, the Pre+2 promoter exhibited some weak constitutive activity in vivo (Table I). This activity presumably results from the 2-base pair insertion which shifts the more consensus-like -35 region sequence (TTGTTT) into the appropriate position as was seen above with the Pre*+2 promoter. This activity was in fact shown to be dependent on the supercoiled integrity of the DNA template, as Pre+2 promoter activity was only observed in vitro with uncut supercoiled template (Fig. 5). Addition of cII protein also had an inhibitory effect on transcription from the Pre+2 promoter, similar to that observed with the Pre*+2 and Pre*-2 promoters (Figs. 4A and 5, Table I). These results support our contention that spatial positioning is crucial for positive activation by cII protein and that changing the spacer length by f 2 base pairs leads to nonproductive cII binding which sterically interferes with the interaction of RNA polymerase a t these promoter sites.

Effects of Alternative DNA Sequences in the -35 Region of
Pre*-The fact that the wild type -35 region sequence of Pre* has little to no homology with the consensus -35 sequence and that neither insertion nor deletion mutations severely effected promoter function suggested that a specific sequence in this region was not critically important for promoter function. T o further test the importance of -35 region sequences on Pre* function we constructed three promoters, Pre*l, Pre*2, and Pre*3, in which the normal -35 region of Pre* was completely replaced with three different DNA sequences (Fig. 2).
The Pre*l construct contained Pre* sequences from position -17 to +2. All other upstream and downstream sequence information was contributed by the pKO-SK vector sequences. Remarkably, this sequence functioned constitutively both in vitro and in vivo despite its poor -35 region sequence information (Fig. 6, Table I). The efficiency of the Pre*l promoter was 60-65% that of the Pre* promoter. This result indicates that although the sequences upstream of position -17 quantitatively affect promoter strength they were cer-  tainly not essential to its function. The two other promoters, Pre*2 and Pre*3, contained essentially "random" DNA sequences inserted upstream of position -19 a t Pre*. Both of these promoters functioned constitutively in the in vitro and in uiuo transcription assays (Fig.  6, Table I). Relative to Pre* activity, Pre*2 was 20-25% as active and Pre*3 activity was approximately 60-65% as active. Again, although quantitative effects were observed these sequences all exhibited quite reasonable levels of promoter activity. This was particularly evident in the case of the Pre*2 promoter which was down approximately 4-5-fold in activity. This promoter contains an extremely poor G/C-rich -35 region sequence (GGGGCG) which to date has never been observed in the -35 region of another constitutive prokaryotic promoter, and yet it retains activity. Thus, although these radically different nonconsensus -35 region sequences can influence Pre* promoter activity, none of the upstream region substitutions interfered with sequences that were critical to promoter function. The results clearly demonstrate that the -35 region sequences can qualitatively affect promoter activity; however, there are no specific sequence requirements in the -35 region of the Pre* promoter which are essential to promoter function.

Point Mutations Upstream of the -IO Region Dramatically
Affect Pre* Function-Our results indicate that at Pres the sequences between positions -16 and +2 are critically important to promoter function. Although the concensus -10 hexamer is certainly a primary determinant of this activity, its presence alone should not be sufficient to create a promoter. We reasoned that at Pre* perhaps sequences surrounding the -10 hexamer were also major contributors to promoter function. Examination of the Pre* sequences immediately upstream of the -10 region hexamer revealed sequence homology with bases that show weak conservation in other promoters (i.e. guanine a t position -14 and thymine at position -15) (8). T o determine the influence of these bases on Pre* function we constructed several derivatives of Pres which contained point mutations a t positions -14, P14T*, P14C* and -15, P15C* (Fig. 2) and tested these Pre* mutants in our in vitro and in uiuo transcription assays.
I n uitro all three of these single point mutations completely eliminated constitutive promoter function (Fig. 7). I n uiuo these mutants also severely affected constitutive promoter activity ( Table I). These results indicate that the thymine and guanine nucleotides present a t positions -15 and -14 of h e * , respectively, are essential for constitutive promoter function. Apparently, in the absence of consensus -35 region sequence information these bases become critical components of the RNA polymerase recognition and initiation signal. We further examined the ability of these mutant Pres promoters to respond to cII activation. In the presence of cII protein, these promoters functioned a t essentially the same efficiency as X Pre (Fig. 7, Table I). Thus, cII can compensate for the deficiency introduced into the Pre* promoter by these position -14 and -15 point mutations just as cII compensates for poor consensus sequence information a t X Pre.

DISCUSSION
We have introduced three single base pair mutations at the X Pre promoter to create a concensus -10 sequence and found that it completely eliminates the cII dependence of this pro- moter. This constitutive promoter, R e * , functions as efficiently as the fully cII-activated X Pre promoter both in uiuo and in uitro. The efficient activity of the Pre* promoter was rather unexpected since its -35 region sequence (GTTTGT) has little to no homology with the consensus -35 region sequence (TTGACA) that is required by most other constitutive prokaryotic promoters for activity. In uiuo and in uitro transcriptional analyses of five Pre* promoter derivatives with alternative -35 region sequences showed that every derivative retained constitutive promoter function. The activity of these promoters was affected by the particular sequences which were placed in the -35 region indicating that they participate in modulating promoter activity. Furthermore, sequences which generally improved the homology of the -35 region with that of consensus (e.g. Pre*+2, TTGTTT) increased promoter efficiency, and those which had poor homology showed somewhat less activity. However, even the Pre*2 promoter which contains a most unusual nonconsensus -35 region sequence (GGGGCG) had only a 4-5-fold down effect on Pre. activity and thus retained quite reasonable levels of promoter activity. These findings clearly demonstrate that the Pre* promoter does not require specific -35 region sequence information for constitutive promoter function, although sequences there may certainly exert a positive or negative effect on overall promoter efficiency.
Our results indicate that the Pre* promoter segment spanning from position -16 to +2 is sufficient for constitutive promoter function. Clearly the consensus -10 region hexamer within this segment is a primary determinant of promoter activity. However, we have identified additional sequences at Pre* just upstream of the -10 region hexamer ( i e . the extended -10 region) which are crucial for constitutive promoter function. Point mutations at either position -14 or -15 dramatically reduce constitutive Pre: activity. Very few mutations have been obtained at analogous sites in other promoter signals. Moreover, the few mutations which have been characterized usually have only minor effects on promoter efficiency (3,10,(26)(27)(28)(29). We suspect that mutations in this extended -10 region of the Pre* promoter eliminate certain contact sites for RNA polymerase and that in the absence of a sufficient -35 region recognition sequence, these contact sites have become critical for promoter recognition and function.
Several lines of evidence support our above contention. First, the extended -10 region does exhibit some weak sequence conservation when all promoter signals are compared (8). However, among those promoters which like Pre* do not require specific consensus -35 region sequence information for activity, such as gut P1, gal P2 (30), omp F (31), and uur A (32), this region, and in particular the G:C and T:A base pairs at postions -14 and -15, respectively, exhibit strict sequence conservation. This data implies that the extended -10 region may actually compensate for the lack of consensus sequence information in the -35 region.
This contention is supported directly by results obtained in mutational studies with the P22 Pant promoter. This promoter has highly conserved consensus sequences in both the -10 and -35 region hexamers but does not contain an extended -10 sequence. Many promoter down mutations have been isolated in Pant and the vast majority mapped to the -10 and -35 region hexamers (9). However, when second site revertants were selected starting with a Pant promoter mutant that contained severe -35 region down mutations, promoter activity was recovered by an A:T to G:C transition at position -14 (33). Hence, the creation of an extended -10 region similar to that defined here at Pre* was able to compensate for the -35 region defect at Pant.
The extended -10 region is also important for efficient promoter function of the X Pre promoter. The Pre mutation cy3019 is a G:C to A:T transition at position -14 which reduces promoter activity (3). Similar point mutations in the extended -10 region of the Pres promoter respond normally to cII activation, and hence an RNA polymerase contact rather than a cII protein contact has been eliminated by the cy3019 mutation. We conclude that the extended -10 region usually plays a relatively minor role in most promoters which contain adequate -35 region sequence information. However, this region does become an important determinant of promoter function in the absence of conserved -35 region sequences.
We also characterized the responsiveness of some of the promoter constructs to activation by cII protein. Transcription from the Pre: promoter was increased about 2-fold in uitro when RNA polymerase was limiting. In the presence of excess RNA polymerase, however, no effect of cII protein was observed. This suggests that cII protein enhances promoter recognition and that this effect can be overcome completely by simply increasing the RNA polymerase concentration. In uitro, again we observed a 2-fold activation of transcription from Pre* by cII protein. This implies that in the bacterial cell the amount of available RNA polymerase is also limiting and that cII functions by helping the Pre* promoter compete for RNA polymerase more effectively.
In those Pre* promoters in which the -35 region had been shifted by either a 2-base pair insertion or deletion mutation we observed inhibition rather than enhancement of transcription by cII protein. In vivo we observed minor reductions in the constitutive activity of these promoters in the presence of cII protein; however, in uitro using saturating concentrations of cII protein a greater degree of inhibition (50%) was observed. Apparently, cII binding actually interferes or competes with RNA polymerase for binding if the cII binding site is displaced within the promoter site. Consistent with this idea was the fact that the identical insertion and deletion mutations at X Pre resulted in complete elimination of cII activation. These results demonstrate that the precise positioning of the cII binding site within the promoter signal is critical for cII-activated transcription from X Pre and the Pre* promoter.