Intrinsic DNA Distortion of the Bacteriophage MumomP1 Promoter Is a Negative Regulator of Its Transcription

The momP1 promoter of the bacteriophage Mu mom operon is an example of a weak promoter. It contains a 19-base pair suboptimal spacer between the −35 (ACCACA) and −10 (TAGAAT) hexamers. Escherichia coli RNA polymerase is unable to bind to momP1 on its own. DNA distortion caused by the presence of a run of six T nucleotides overlapping the 5′ end of the −10 element might prevent RNA polymerase from binding to momP1. To investigate the influence of the T6 run on momP1 expression, defined substitution mutations were introduced by site-directed mutagenesis. In vitro probing experiments with copper phenanthroline ((OP)2Cu) and DNase I revealed distinct differences in cleavage patterns among the various mutants; in addition, compared with the wild type, the mutants showed an increase (variable) in momP1 promoter activity in vivo. Promoter strength analyses were in agreement with the ability of these mutants to form open complexes as well as to producemomP1-specific transcripts. No significant role is attributed to the overlapping and divergently organized promoter,momP2, in the expression of momP1 activity, as determined by promoter disruption analysis. These data support the view that an intrinsic DNA distortion in the spacer region ofmomP1 acts in cis as a negative element inmom operon transcription. This is a novel mechanism of regulation of toxic gene expression.

Optimal activity of bacterial promoters depends on the precise and controlled interactions between the promoter with regulatory proteins and RNA polymerase (RNAP). 1 In many instances, promoter activity is modulated by protein-induced changes in DNA structure such as DNA distortion, looping, bending, and unwinding (1)(2)(3). DNA structural distortions are known to influence promoter activity (4,5). Certain oligo (A/T) tracts exhibit unusual curvature (6) and play an important role in the regulation of transcription initiation (7)(8)(9)(10)(11). A different role is attributed to the A tract when it is positioned upstream and in phase with the promoter elements. A number of Esche-richia coli promoters have AϩT-rich tracts (also known as UP elements) upstream to the Ϫ35 hexamer of 70 promoters (12,13). UP elements, when present, are integral components of promoters, because they interact with the carboxy terminal domain of the RNAP ␣-subunit (12). In vitro studies showed that the E. coli RNAP ( 70 ) holoenzyme alone is sufficient for transcriptional activity from several such promoters (12,14,15).
The regulatory region of the mom operon of bacteriophage Mu, which controls a unique DNA modification function (see Ref. 16 for a recent review), exhibits several interesting features. The promoter, momP1, which directs the transcription of com-mom dicistronic mRNA, is a typical example of a weak promoter with a poor Ϫ35 (ACCACA) element and a suboptimal spacing of 19 bp between the two consensus elements (Fig.  1a). The spacer region of the promoter contains a run of six T nucleotides from Ϫ12 to Ϫ17. RNAP does not bind to momP1 by itself (17). Instead it binds an overlapping, divergent promoter region, momP2, which brings about "leftward" transcription (18). The stretch of six A nucleotides complementary to the T 6 run appears to be part of an UP element for leftward transcription from momP2 (18).
The regulation of mom operon expression occurs at both the transcriptional and translational levels (16). The Mu C protein, a "middle" gene product, is an obligatory transcriptional activator of the momP1 promoter (19,20), as well as for the other three late promoters (21,22). C protein binding to a site located at Ϫ28 to Ϫ57 in the momP1 region (23,24) brings about an asymmetric distortion and unwinding of the DNA (25)(26)(27).
Mutants have been isolated that relieve the dependence of momP1 on C activation. In one such (partially) C-independent mutant, tin7, there is a single-base change (T to G at position Ϫ14) that disrupts the T 6 run in the momP1 spacer region (17). Few explanations could account for the increased promoter activity of tin7 mutant: 1) T tract-mediated intrinsic DNA curvature is lost because of disruption of the T 6 run (to T 3 GT 2 ), resulting in RNAP binding to momP1; 2) because P1 and P2 are overlapping divergent promoters, weakening of the UP element of momP2 (18) may facilitate the binding and activity of RNAP at momP1; and 3) the T(Ϫ14) to G change converts momP1 to an extended Ϫ10 promoter (28). The present study is an attempt to delineate the role of the T 6 run in momP1 promoter activity. zymes were purchased from Stratagene, New England Biolabs, and Roche Molecular Biochemicals and were used according to the suppliers' recommendations. DNase I was from Worthington, and E. coli DNA polymerase (PolIk) was from New England Biolabs. Superscript reverse transcriptase was purchased from Life Technologies, Inc. Chemicals and other reagents were purchased from Life Technologies, Inc. and Sigma. Primers were synthesized by Bangalore Genei (Pvt.) Ltd. (Bangalore, India), Life Technologies, Inc., and the University of Rochester Core Lab Facility. [␥-32 P]ATP (6000 Ci/mmol) was purchased from New England Nuclear. Most of the standard procedures were carried out as described by Sambrook et al. (30).
Construction of momP1 and momP2 Mutants-The mutants used in this study were generated by site-directed mutagenesis using either pLW4 (17) or pUW4 (31) as the template DNA. pUW4 was used as template for the polymerase chain reaction-based mutagenesis methods. The mutants, pT2G, pT3G, and pT2GT3G (Fig. 1b), were generated by using the Stratagene QuickChange TM site-directed mutagenesis method involving a pair of mutagenic oligonucleotides and PfuI DNA polymerase. The mutant pT3C was generated by using a Promega Gene Editor site-directed mutagenesis kit. pT1C, pT5C, pT4C, pWT-P2, ptin7-P2, pT2GT3G-P2, and pG21C were generated by using the modified mega primer method. In this method, a mega primer was first generated using the pUC reverse primer and the mutagenic oligonucleotide (as described in Ref. 32). The mega primer was then used in the Stratagene QuickChange TM site-directed mutagenesis method. All of the mutants generated in the pUW4 background were subcloned into pLW4 using EcoRI and BamHI restriction enzymes to generate the promoter mutants as lacZ transcriptional fusions. All of the mutants generated were confirmed by carrying out Sanger's dideoxy method of sequencing (30).
The promoter expression plasmid pLC1 (22) was generously provided by Dr. M. M. Howe; it contains an EcoRI-SmaI-BamHI linker upstream of a promoterless lacZ gene. Plasmid pLO1 was created by cloning the smaller PstI-BamHI fragment from pLC1 into pRSGC3 SmaI (a derivative of phagemid pGC1; Ref. 33). A synthetic duplex containing either momP1 or momP2 was generated by annealing pairs of appropriate synthetic oligodeoxynucleotides that had appropriately located 5Ј EcoRI and 5Ј BamHI single-strand overhangs. Plasmids pLO1/P1 and pLO1/P2 were constructed by ligating the synthetic duplexes into the EcoRI and BamHI sites, respectively, of pLO1 and were used for generating site-directed mutations in momP1 and momP2, respectively. After DNA sequencing confirmed the nature of each mutation, the mom promoter-containing PstI-BamHI fragment was cloned into the corresponding sites in pLC1 for promoter expression analyses (additional details of the plasmid and mutant constructions are available upon request).
Promoter Strength Analysis-Isolated colonies of E. coli DH10B cells harboring either a promoter mutant plasmid alone or with plasmid pVN184 were inoculated into LB broth containing 100 g/ml of ampicillin (for mutant promoter plasmids alone) or ampicillin and 25 g/ml chloramphenicol (both plasmids present); the cultures were incubated at 37°C for ϳ16 h with vigorous shaking. The overnight cultures were diluted 100-fold into 3 ml of fresh medium in duplicate tubes and incubated at 37°C till the cultures reached an A 600 of 0.3-0.7. The samples were then placed on ice. ␤-Galactosidase activity in SDS-CHCl 3 -treated cells was determined as described by Miller (34). In the experiments with plasmid pLC1 constructs, ␤-galactosidase assays were carried out with exponential cultures grown from isolated colonies. The values in the tables are the averages from at least two separate experiments, and replicate assays were done on each culture. The variation was 10 -20% around the mean value.
DNase I and (OP) 2 Cu Cleavage Reactions-2.0 g (ϳ0.36 pmol) of negatively supercoiled DNA was incubated with DNase I (final concentration, 0.1 ng/l) in presence of buffer (20 mM Tris-HCl, pH 7.2, 1 mM EDTA, 5 mM MgCl 2 , and 50 mM NaCl) in a total reaction volume of 20 l. After 30 s at 22°C, the reaction was terminated by addition of 20 l of stop buffer (0.1 M Tris-HCl, pH 7.5, 25 mM EDTA, and 0.5% SDS). The sample volume was made to be 400 l with water and extracted successively with phenol/chloroform/isoamyl alcohol (25:24:1) and chloroform/isoamyl alcohol (24:1) and then precipitated with 2.5 volumes of 100% (v/v) ethanol in the presence of glycogen as a carrier. Primer extension protocol is adapted from Gralla (35). The extension reactions were performed with end-labeled mom forward and reverse primers as previously described (27).
In Vivo KMnO 4 Footprinting Reaction-In vivo KMnO 4 footprinting reaction was carried out as described by Sasse-Dwight and Gralla (36). E. coli DH10B cells harboring a momP1 promoter mutant plasmid alone or along with pVN184 were grown to A 600 0.6 in 4.0 ml of LB broth. The cultures were treated with 200 g/ml rifampicin for 20 min. The samples were then incubated with 30 mM KMnO 4 for 2 min. Reactions were stopped by transferring the cultures to prechilled tubes. The cells were harvested, and plasmid DNA was isolated. Primer extension reactions were carried out as described above.
Total RNA Isolation and Primer Extension-Total RNA was isolated from E. coli DH10B cells harboring the various promoter mutant plasmids using the hot acid phenol method. Primer extension was carried out as per the manufacturer's protocol (Life Technologies, Inc.) using superscript reverse transcriptase and end-labeled mom forward (for momP2 transcript detection) and reverse (for momP1 transcript detection) primers. An end-labeled primer annealing 150 bases downstream of the ampicillin transcription ϩ1 start site was used to normalize the levels of transcripts produced in the different mutant promoter constructs. Scanning of the autoradiographs was carried out using a Bio-Rad GS710 Calibrated Imaging Densitometer. Quantification was done using Quantity One software.

DNA Structure Analysis of T 6 Run Mutants-The variation
in helical structure of the DNA depends on base sequence. Specific sequences contribute to alterations in groove width and DNA curvature (37). In addition to their use in probing DNA-protein interactions, nucleases are often used to detect distortions in DNA. Cleavage reaction of orthophenanthroline cuprous complex ((OP) 2 Cu) depends on the local DNA structure rather than the base sequences as demonstrated previously by Spassky et al. (38). We used (OP) 2 Cu to probe possible structural or conformational differences between the wild type (WT) and the tin7 mutant in the region of the T 6 tract in momP1 ( Fig.  1). Negatively supercoiled plasmid DNA harboring the WT or tin7 mutant momP1 promoter was subjected to in vitro single hit cleavage, and the sensitivity pattern was assessed by primer extension analysis (Experimental Procedures). The results of a typical (OP) 2 Cu footprinting are shown in Fig. 2 (a and b). The sensitivity patterns of both the top and the bottom strands in the region containing the T 6 run were different for the two promoters. Several hypersensitive sites are seen in the WT that are not reactive in tin7. For example, at Ϫ14T (top strand) and at Ϫ15A, Ϫ16A, Ϫ17A, and Ϫ18C (bottom strand), the WT was cleaved more often by (OP) 2 Cu. In contrast, tin7 DNA was relatively refractory to cleavage by (OP) 2 Cu at these residues, whereas Ϫ10A in the top strand was hypersensitive. We also probed the promoter structure by using DNase I as a footprinting agent. DNase I reaction also revealed substantial differences in cleavage sensitivity patterns (indicated by the asterisks in Fig. 2c, lanes 1 and 2). DNase I cleavage gave rise to two hypersensitive sites, at Ϫ9G and Ϫ17T (top strand) in the WT promoter, compared with hypersensitive sites at Ϫ12T and Ϫ13T (in the T 6 run) of the tin7 promoter. These results show that the two promoter regions differ in their susceptibility to nuclease cleavage, indicating that the DNA conformations are different.
Because the T4G (tin7) mutant showed a difference in DNA conformation with respect to wild type, the effect of base substitutions at other position in the T 6 run were examined. To this end, negatively supercoiled DNAs of various mutants were subjected to in vitro cleavage with DNase I. Mutants T2G, T3G, T2GT3G, and T4G (tin7) were selected as representatives for this analysis (Fig. 2c). The mutants showed hypersensitivity patterns different from one another, as well as from the WT. For example, the top strand residue Ϫ14T was cleaved more frequently in T2G, whereas Ϫ12T, Ϫ14T, and Ϫ16T were hypersensitive in T3G, and Ϫ12T and Ϫ14T were hypersensitive in T2GT3G. The narrower minor groove of the A/T tract is altered by the G substitutions, leading to its widening at these positions. As a consequence, different DNase I-hypersensitive sites are observed (marked by asterisks in Fig. 2c) in each of the mutants; (OP) 2 Cu footprinting analysis gave analogous results (not shown). Because both DNase I and (OP) 2 Cu-mediated DNA cleavages are in the minor groove (39,40), the T 6 run mutations appear to change the DNA conformation in the minor groove. This is further supported by an altered migration pattern of DNA fragments in polyacrylamide gels (data not shown).
Mutations in the T 6 Run Result in Increased Promoter Activity-The differences in chemical nuclease and enzymatic cleavage patterns of WT and mutant T 6 run promoters reflect structural differences among these promoters. To determine whether the changes also influence promoter activity, promoter-lacZ fusion constructs were generated for all of the mutants, and promoter strength was assessed indirectly by measuring ␤-galactosidase activity in cells harboring these plasmids. Moreover, the C-activated level was analyzed in cells also harboring a compatible C-producing plasmid, pVN184. T 6 run mutants T2G, T2C, T3G, T3C, T2GT3G, and T5C produced 5-26-fold higher levels of enzyme compared with the WT promoter (Table I); in contrast, T1C and T4C showed little or no increased expression. However, all of the mutants remained responsive to transactivation by C protein (Table I), producing enzyme levels comparable with that of the activated WT momP1 promoter. Thus, all of the mutant promoters are C protein-dependent for their full activity, indicating that the C transactivation mechanism was unaltered. As a control, the mutation G21C was created (shown in Fig. 1b) upstream from the T 6 run yet within the spacer region. As expected, this mutant showed levels of ␤-galactosidase activity comparable with the WT momP1 promoter in the absence and in the presence of C protein (Table I). The variable increase in momP1 promoter activity among these mutants could be due to differences in their perturbations of DNA structure as shown in Fig.   2. None of the mutant promoters showed activity as high as that of T4G (tin7), which showed an increase that was between 46-and 80-fold depending on the type of fusion examined (Tables I and II). These results suggest that in addition to DNA distortion, an alternative mechanism might be operating in tin7, most likely having an extended Ϫ10 promoter because of the specific base substitution at Ϫ14 position (discussed further below).
The above experiments were carried out using a momP1 promoter directing production of a Com-LacZ translational fusion. We carried out similar experiments with a momP1 promoter-lacZ transcriptional fusion vector. This was constructed by subcloning momP1 mutations (T4G, T4A, T4C, T3G, T3A,  and T3C, produced in a momP1-containing synthetic oligonucleotide duplex) into a site 5Ј to a promoterless, reporter lacZ gene (see "Experimental Procedures"), and momP1 promoter activity was assayed by measuring ␤-galactosidase activity. As seen in Table II, the substitutions generated variable increases in enzyme level, in good agreement with the results observed with the pLW4 plasmid system. Most interesting are the three T4X substitutions. First, the T4G mutant had the highest level of C-independent expression, 80-fold above the WT. In contrast, T4A had a 6-fold increase, whereas T4C showed no increase. Thus, the three different T4X substitutions produced three different phenotypes. We suggest that the high level of constitutive expression by T4G (tin7) is due to its having an extended Ϫ10 promoter, in addition to the alteration in DNA conformation. In contrast, the T4A (as well as the T3A) substitution appears to only affect momP1 DNA conformation, indicating that T-A to A-T base pair alterations can also affect conformation. At first glance, it was surprising that the T4C mutant did not show increased momP1 expression; however, as will be shown below, the T4C mutant does not exhibit any structural difference from the WT based on in vitro cleavage. Finally, it should be noted that the T2G and T3G mutations create a TG at positions Ϫ17 and Ϫ16 and at positions Ϫ16 and Ϫ15, respectively. Although these mutations exhibited enzyme levels severalfold higher than the WT, they do not appear to provide extended Ϫ10 functional capability. In view of the different levels of momP1 expression observed with the three T4X mutations, we compared their in vitro sensitivity to cleavage by DNase I. As shown in Fig. 2d, the T4C promoter region showed a cleavage pattern similar to that of the WT, which is in sharp contrast to all of the other T 6 run mutants. Because the T4C substitution does not appear to alter the WT momP1 DNA conformation, we suggest that it possesses the same unfavorable distortion as the WT and, hence, requires protein C activation for any transcription.
Formation of Open Complexes by T 6 Run Mutants-Increased ␤-galactosidase levels with certain T 6 mutant momP1 promoters indicate increased transcription initiation capability. Interaction of RNAP at a promoter can be ascertained by assessing open complex formation using an in vivo KMnO 4 footprinting technique (36). E. coli DH10B cells harboring a T 6 mutant pLW4 plasmid in the absence or the presence of the C protein-producing plasmid, pVN184, were probed (see "Exper-imental Procedures"). In accordance with its high level of constitutive promoter activity in the absence of C protein, tin7 showed hypersensitive bands (Fig. 3, lane 4) characteristic of open complex formation; the observed pattern was in good agreement with the results of Balke et al. (17). Open complex formation in the absence of C protein was also observed with mutants T2G and T2GT3G (Figs. 3, lane 12, and 4a, lane 10). However, open complexes were not detected with T3G, T3C, and T2C (Fig. 3, lanes 8, 10, and 14, respectively), which correlates with their relatively lower levels of ␤-galactosidase expression (in the absence of C protein). As could be predicted from the promoter strength analysis (Table I), the WT momP1 promoter and the T4C mutant were unable to produce detectable open complexes (Fig. 3, lanes 2 and 6, respectively). In the presence of C protein, however, all promoters showed open complex formation (Fig. 4, a and b). This result rules out an artifactual inability to detect open complexes with the mutant constructs used in these experiments.
T 6 Run Mutants Show Increased P1 Transcript Levels-Because only mutants with higher (Ͼ8-fold the WT basal level) expression of ␤-galactosidase activity showed open complex formation, we employed a more direct method of assessing promoter strength. For this purpose, we assayed for momP1specific mRNA transcripts in total RNA isolated from WT, T4G (tin7), and T 6 mutant (T2G, T3G, T3C, and T2GT3G) plasmidcontaining cells (see "Experimental Procedures"). The results of such an experiment are shown in Fig. 5a. In all of the mutants examined the transcription start site was identical to that of the wild type momP1 promoter, indicating that the mutations did not lead to the formation of new promoters. Those mutants (e.g. T3G and T3C) that failed to show open complexes in the KMnO 4 probing experiments did produce increased amounts of momP1-specific transcripts compared with the WT promoter (Fig. 5). There was a good correlation in the fold increase in momP1-specific transcript levels and the relative promoter strengths of these mutants with respect to the WT levels (compare Tables I and II and Fig. 5b).
Mutations Disrupting the momP2 Ϫ10 Hexamer Do Not Increase Activity of the WT (or T 6 Run Mutant) momP1 Promoter-The results presented above support the view that alteration in DNA conformation caused by disruption of the T 6 run results in increased basal activity of the momP1 promoter. However, the scenario is somewhat complicated by the fact that the mom regulatory region has two overlapping divergent promoters, momP1 and momP2. The T 6 run substitution mutations generated in momP1 also disrupt the A 6 tract in the complementary strand (Fig. 1), which is proposed to function as part of an UP element directing leftward transcription from the momP2 promoter (18). Hence, an alternate possibility for the increased activity of momP1 in tin7 and other T 6 run mutants could be due to weakening of the UP element of momP2. Therefore, momP2 transcript levels were measured by isolating total RNA and extending it with end-labeled mom forward primer using reverse transcriptase. The results are shown in Fig. 6, where momP2 transcripts were detected with tin7, as well as with some other T 6 (T2G, T3G, and T3C) run mutants. Thus, the substitution mutations in the T 6 /A 6 run did not abolish momP2 activity while having increased momP1 activity. This conclusion was supported by results from independent experiments in which synthetic duplexes having mutations in the T 6 /A 6 run (corresponding to T2G or T3G or T4G) were cloned into pLC1 in an orientation where lacZ gene transcription was under control of momP2 (in these constructs the momP1 Ϫ10 hexamer was also altered so as to reduce its potential expression). We observed that each of the single-base substitution mutations lowered the enzyme level less than 2-fold (data not shown). Thus, T 6 run mutations that increase momP1 transcription do not do so by reducing momP2 transcription.
To further examine the effect (if any) of momP2 expression on momP1 expression, the momP2 Ϫ10 hexamer was mutated in the WT, tin7, and T2GT3G mutant constructs (Fig. 7a). In these mutants, loss of momP2 function was confirmed by measuring leftward transcript levels produced in vivo by the parental and disrupted momP2 promoters, using primer extension assays with total RNA extracted from cells harboring these plasmids (Fig. 7b). The results in Fig. 7c show that there was no increase in the WT, tin7, or T2GT3G momP1 promoter activity in the momP2 Ϫ10 disrupted background. These results indicate that the overlapping momP2 promoter plays, at most, only a minor role in momP1 activity, unlike other overlapping promoters. We conclude that mutations in the T 6 run that increase momP1 expression function by alleviating DNA distortion. DISCUSSION We have addressed the importance of the run of six T nucleotides located in the momP1 promoter (Fig. 1a) in the regulation of mom operon expression. An intrinsic DNA distortion caused by the presence of the T 6 tract overlapping the 5Ј end of the Ϫ10 element could produce an unfavorable conformation for RNAP occupancy. Different T to G substitutions in this run showed different sensitivity patterns to nucleases as compared a The relative ␤-galactosidase activity with the WT promoter in the absence of C protein is defined as 1.0. It corresponds to 23 Miller activity units.
b Fold activation is defined as the ratio of ␤-galactosidase activity produced by a momP1 mutant plasmid in the presence versus the absence of the compatible C protein-producing plasmid, pVN184. a The relative ␤-galactosidase activity with the WT promoter is defined as 1.0. It corresponds to 20 Miller activity units. with the WT momP1 promoter (Fig. 2). This is attributed to a difference in the DNA structure caused by each substitution. These substitutions in the T 6 run also produce variable increases in the basal activity of mutant momP1 promoters (Tables I and II). The increase in the promoter strength of some of the mutants was correlated with the formation of detectable open complexes and the levels of momP1-specific transcripts in the absence of trans-activator protein C (Figs. 3-5).
One could argue that the increase in promoter activity observed in the T 6 run mutants could be a consequence of new base-specific contacts made by RNAP in the substituted positions instead of removal of intrinsic DNA distortion. However, this seems unlikely because changing different residues (Ϫ13 to Ϫ16) led to the increase in promoter activity. Moreover, substitutions elsewhere in the spacer do not increase the activity, including the mutant T1C, whose substitution still retains the run of T 5 nucleotides. A point to be remembered and shown here is that momP1 wild type promoter is unable to form an open complex on its own (Ref. 17 and Fig. 3). This is not due to the suboptimal spacer length (see the Introduction and Fig.  1) because single-base and two-base deletions in the spacer (18 and 17 bp, respectively) do not lead to any increase in promoter activity (data not shown). Thus, it is difficult to visualize RNAP contacting each and every residue between Ϫ13 to Ϫ16 when it is unable to make favorable contacts in an optimally spaced promoter. An altered DNA structure of the various T 6 run mutants leads to the formation of an open complex at these promoters. As an additional support for altered DNA conformation detected in nuclease probing experiments, the wild type promoter fragment (226 bp) migrated slower than the T2G mutant promoter fragment (226 bp) in a gel electrophoretic mobility assay. Taken together, we conclude that the transcription from promoters having substitutions in the T 6 run is due to the removal of an unfavorable distortion.
Of all of the mutants whose promoter strength we analyzed, T4G (tin7) showed the highest momP1 transcriptional activity. This mutation, a T 3 G at position Ϫ14, produces a Ϫ15T, Ϫ14G, which is characteristic of extended Ϫ10 promoters (28,41). Extended Ϫ10 promoters are usually constitutive, and they do not require a Ϫ35 element or an activator protein. In contrast to the T4G, the corresponding T4C substitution did not increase expression of momP1 nor alter promoter DNA conformation compared with the WT. All three Ϫ15T (T3X) substitutions exhibited an increase in momP1 basal activity compared with WT. However, although the T3G mutation created a TG at positions Ϫ16 and Ϫ15, it did not produce the same high level of expression exhibited by T4G (tin7); this is to be expected because the former TG is not positioned properly to create an extended Ϫ10 promoter. Thus, we suggest that a combination of both DNA conformational alteration and extended Ϫ10 promoter characteristics contribute to the T4G (tin7) phenotype, but the increase in basal activity of the other T 6 run mutants is due to the removal of an unfavorable distortion in momP1 promoter DNA. Once this distortion is ameliorated, an otherwise very weak promoter can be transcribed in the absence of activator protein C. However, these promoters are still dependent on C for full activity, as shown by both the C-mediated increases in ␤-galactosidase activity (Table I) and the formation of open complexes (Fig. 4).
It has been shown earlier that it is primarily the length, not the sequence, of spacer DNA between the two promoter consensus sequences (Ϫ10 and Ϫ35 regions) that is important for activity of a promoter (42,43). It has also been demonstrated that the sequences located either upstream or downstream of the Ϫ10 and Ϫ35 regions determine the kinetics of association of promoter with RNAP and efficiency of transcription initiation (44,45). It is believed that the spacer DNA holds the Ϫ10 and Ϫ35 regions in the proper orientation for their recognition by the RNAP holoenzyme complex without having any specific contacts with RNAP. However, characterization of mutants of the P RM promoter of phage bearing dC 9 ⅐dG 9 sequences in a stretch of the spacer DNA separating the contacted Ϫ10 and Ϫ35 regions showed reduced promoter activity both in vitro and in vivo (46,47). These mutations were interpreted as altering the structure of the spacer DNA and, as a consequence, leading to a change in the orientation or local structure of the contacted Ϫ10 and Ϫ35 elements of the promoter. A library of synthetic promoters of Lactococcus lactis having randomized 17-bp optimal length spacer in between the consensus Ϫ10 and Ϫ35 elements was assayed for activity both in L. lactis and E. coli (48). In both host backgrounds, a large variation (ϳ400 fold) in promoter activity was observed because of variations in the spacer sequence context. It seems that the overall threedimensional topological structure of the promoter DNA that arises from a particular nucleotide sequence could be important for the activity of a promoter.
Complex regulatory mechanisms have evolved in bacteriophages to ensure the precise expression of phage genes. Expression of the bacteriophage Mu mom gene during the late The experiment was carried out as described under "Experimental Procedures" using equal amounts of total RNA (20 g) in all of the lanes and end-labeled mom forward primer for primer extension. transcription phase is a good example of one such regulatory scheme. Although mom gene expression seems to be dispensable for Mu growth or lysogeny, premature activation or constitutive expression of mom function is detrimental to the host cells (49,50). Hence, it is not surprising that intricate mechanisms have been evolved for the regulation of mom expression at both the transcriptional and translational levels (16). Apart from these modes of regulation, recently, Sun and Hattman (18) have suggested another possible regulatory control over mom expression. It was suggested that the leftward transcription at the momP2 promoter ( Fig. 1 and Ref. 18) might prevent low level rightward transcription of momP1 (and, hence, mom expression) in two possible ways. First, momP2 might compete with momP1 for RNAP binding in the absence of C protein. Second, leftward transcription produces an antisense transcript that might prevent gin mRNA elongation into mom. The present study rules out the first possibility because disruption of the momP2 Ϫ10 hexamer did not lead to an increase in the basal level activity of momP1 (Fig. 7c). On the other hand, disruption of the T 6 run in the spacer region of momP1 promoter led to increased rightward transcription, indicating the importance of its role as a cis-acting negative element. Another possible role attributed to momP2 is to act as a sink for capturing RNAP in the vicinity of momP1, so that RNAP is ready for occupancy at momP1 at the right time of mom expression (17). Our results do not exclude that possibility.
Existence of overlapping promoters in many systems add additional regulatory complexities (51)(52)(53). The momP1 and momP2 promoters (Fig. 1a) are overlapping and oriented in a divergent fashion. Normally such organization would lead to competition in the transcription machinery as exemplified in case of other overlapping/competing promoters (53). However, earlier DNase I footprinting analysis with the mom promoter revealed that RNAP is bound to momP2 in the absence of C protein (17). Partitioning of RNAP between the two promoters was not observed, although neither promoter appears to be a strong one. Further, momP2 disruption did not lead to increased momP1 expression (Fig. 7c), underlining the importance of the T 6 run as a cis-acting negative element that prevents RNAP binding to momP1. Thus, the primary role of the T 6 run is to prevent low level rightward transcription initiation at momP1.
It is noteworthy that the P lys promoter, another bacteriophage Mu late gene promoter, also has a T 6 stretch in the position corresponding to that in momP1, but it is absent in the other two late promoters (P i and P p ). Substitutions in two of the T bases in the P lys spacer region show an UP phenotype depending on the substituted base (22). Because the premature expression of mom and lys is detrimental to the host cells, the DNA negative element seems to be a common fail-safe mecha- nism to keep these two late genes tightly regulated until the right time for their expression. Thus, the phage Mu seems to have evolved one common strategy to keep two potentially cytotoxic genes under control.