Recognition of the -10 Promoter Sequence by a Partial Polypeptide of sigma 70 in Vitro*

(Received for publication, October 9, 1996, and in revised form, November 29, 1996)

Alicia J. Dombroski Dagger

From the Department of Microbiology and Molecular Genetics, University of Texas Health Science Center, Houston, Texas 77030

ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
FOOTNOTES
Acknowledgments
REFERENCES


ABSTRACT

Promoter recognition by RNA polymerase depends upon its ability to bind to specific DNA sequences. The sigma (sigma ) subunit provides selectivity for transcription initiation by interacting with the -10 and -35 elements of promoter DNA. Suppressor mutations in sigma  factor that compensate for specific "down" substitutions in the promoter have demonstrated that sigma  factor recognizes certain base pairs of the promoter. Since these suppressors were only identified for changes at the -12 and -11 positions of the -10 element (<UNL>TA</UNL>TAAT), the role of the other base pairs of this region in specifying recognition by sigma  factor remained unclear. Using a partial polypeptide of sigma 70 carrying regions 2-4, this report shows that the first three positions of the -10 element (-12, -11, -10) are important for sigma  factor alone to recognize and bind to duplex DNA. The sigma  polypeptide also binds to an "extended -10" promoter, even without a -35 element. A mismatch bubble from -10 to +3 is bound regardless of the sequence within the bubble, or the presence or absence of a -35 element. Unexpectedly, binding to a mismatch bubble that lacks a -35 hexamer is sensitive to the identity of the -11 position, but not the -12 position.


INTRODUCTION

Initiation of transcription at bacterial promoters requires the sigma (sigma ) subunit of RNA polymerase (1, 2). sigma  factor directs holoenzyme (alpha 2beta beta 'sigma ) to the recognition sequences upstream of the transcription start site, and it is required for strand opening on circular DNA (3). A variety of different sigma  subunits is available for positive regulation of gene expression in response to environmental stimuli (4). Each sigma  subunit recognizes a distinct DNA sequence and thereby confers selectivity for initiation upon RNA polymerase.

The sigma  subunits are evolutionarily and functionally related. The largest family comprises those most similar to the primary sigma  factor from E. coli, sigma 70 (5). Members of this family utilize promoters that comprise two hexameric DNA sequences at approximately -35 and -10 relative to the start point of transcription (6, 7). For sigma 70, the consensus promoter consists of TTGACA centered at -35 separated by 17 base pairs from TATAAT centered at -10. Deviations from consensus usually result in reduced promoter activity (8, 9). One exception is the "extended -10" promoter, where an additional TGn is found immediately preceding the -10 hexamer, and a good match to the -35 consensus is not required for activity (10-15).

During initiation, base pairing of the promoter in the region from approximately -10 to +2 is disrupted and a transcriptionally competent open complex is formed (16-19). The precise role of sigma factor in this process is not clear, but sigma  factor cross-links between -10 and +1 on the non-template strand (20, 21). Recently, sigma 70 holoenzyme has been shown to recognize the non-template strand of the -10 sequence in a melted transcription bubble (22). Others have proposed that sigma  factor stabilizes the single-stranded DNA in the open complex by interacting with chemical groups in the non-template strand (23).

Amino acid sequence analysis has revealed four highly conserved regions within the sigma 70 family (Fig. 1; Refs. 4, 5, and 24-26). Region 1.1 inhibits DNA binding by the C-terminal DNA binding domains of sigma 70 in the absence of the core subunits (alpha 2beta beta '), and may physically interact with amino acids in or near the region 4 helix-turn helix domain at the C terminus (27, 28). We have recently shown that holoenzyme containing a sigma 70 derivative, which lacks Region 1.1, is significantly slower in progressing from the initial closed complex to a strand-separated initially transcribing complex. Additionally, if both Regions 1.1 and 1.2 are missing from sigma 70, holoenzyme is arrested in the first closed complex.1 Region 2.4 recognizes the -10 hexamer (29-33), and recent studies suggest that region 2.3 may participate in the DNA melting stage of initiation (34, 35). Region 4.2 is involved in recognition of the -35 hexamer (27, 29, 36), while region 4.1 has been proposed to contact certain transcriptional activators during initiation (37-39).


Fig. 1. Linear diagram of sigma 70. The most highly conserved blocks of amino acid sequence among the sigma 70 family of sigmas are numbered 1-4 (5). Subregions are indicated by shaded regions within the boxes. The core binding domain is designated cb (58). Region 1.1 inhibits DNA binding by sigma  in the absence of the core subunits (27, 28). Region 2.3 is implicated in promoter melting during initiation (34). Region 2.4 interacts with the -10 promoter consensus. Region 4.2 contains a helix-turn-helix motif and interacts with the -35 promoter consensus.
[View Larger Version of this Image (8K GIF file)]


The initial assignment of DNA recognition regions to the sigma  polypeptide was accomplished using genetic analyses, starting with point mutations in the -10 or -35 hexamers that severely reduced promoter activity. Suppressor mutations in sigma  factor, which improved utilization of promoter "down" mutations in the -10 hexamer, localized to region 2.4 (29-32). Interestingly, mutations were only obtained for suppression of the first two positions of the hexamer. There are several possible explanations for this observation. The sigma subunit may only interact with the first two positions of the hexamer, or the search for suppressor mutations was not saturating, or mutations that affect recognition of the +1 proximal positions are lethal. Thus, in this report a direct evaluation of how sigma 70 interacts with the -10 sequences of the promoter was conducted.

Using a partial polypeptide of sigma 70 and derivatives of the tac promoter, several aspects of promoter recognition by sigma  in the absence of the core subunits of RNA polymerase have been addressed. First, the importance of each position in the -10 hexamer was evaluated. Second, the requirements for interaction with an extended -10 promoter were examined in the presence and absence of a -35 hexamer. Finally, the role of single-stranded DNA was assessed both alone, and as a component of a mismatch bubble.


EXPERIMENTAL PROCEDURES

Materials

Restriction endonucleases, T4 DNA polymerase, T4 DNA ligase, and T4 polynucleotide kinase were from New England Biolabs or Promega. Taq DNA polymerase was from Perkin Elmer or Fisher Scientific. [gamma -32P]ATP (3000 Ci/mmol) was from Amersham Corp. Nitrocellulose filter discs (24 mm) were from Schleicher & Schuell or Millipore. Oligonucleotides were synthesized by either Genosys or Bioserve Biotechnology.2

Construction of Promoter Mutations

Single-stranded M13mp19ptac (27) was used as the template for site-directed mutagenesis. Synthetic oligonucleotides (15-10-mers) were used according to Kunkel et al. (40) to generate the point mutations at -7 to -12 and -32, as well as the extended -10 version of the tac promoter as shown in Table I. The mutated M13 derivatives were used as templates for polymerase chain reaction (PCR)3 amplification of DNA to be used for competition nitrocellulose filter binding. Non-promoter DNA (Delta P) consisted of sequences from the M13 polylinker cloning region.

Table I.

Variants of the tac promoter for DNA binding

Sequences shown are contained within a 100-base pair fragment of DNA. Changes made within or near -10 hexamer are underlined and are compared to the ptac sequence at the top of the figure. Letters in bold indicate the -10, extended -10, or -35 consensus sequences.
                         -35                        -10       
ptac: 5'-...TGAAATGAGCTGTTGACAATTAATCATCGGCTCGTATAATGTGTGGAATTG...-3'
ptac-7G: 5'-...TGAAATGAGCTGTTGACAATTAATCATCGGCTCGTATAA<UNL>G</UNL>GTGTGGAATTG...-3
ptac-8G: 5'-...TGAAATGAGCTGTTGACAATTAATCATCGGCTCGTATA<UNL>G</UNL>TGTGTGGAATTG...-3'
ptac-9G: 5'-...TGAAATGAGCTGTTGACAATTAATCATCGGCTCGTAT<UNL>G</UNL>ATGTGTGGAATTG...-3'
ptac-10G: 5'-...TGAAATGAGCTGTTGACAATTAATCATCGGCTCGTA<UNL>G</UNL>AATGTGTGGAATTG...-3'
ptac-11C: 5'-...TGAAATGAGCTGTTGACAATTAATCATCGGCTCGT<UNL>C</UNL>TAATGTGTGGAATTG...-3'
ptac-12C: 5'-...TGAAATGAGCTGTTGACAATTAATCATCGGCTCG<UNL><B>C</B></UNL>ATAATGTGTGGAATTG...-3'
ptac-12G: 5'-...TGAAATGAGCTGTTGACAATTAATCATCGGCTCG<UNL><B>G</B></UNL>ATAATGTGTGGAATTG...-3'
ptac-32A: 5'-...TGAAATGAGCTGTT<UNL>A</UNL>ACAATTAATCATCGGCTCGTATAATGTGTGGAATTG...-3'
ptac-ext-10: 5'-...TGAAATGAGCTGTTGACAATTAATCATCGGCT<UNL>GC</UNL>TATAATGTGTGGAATTG...-3'
ptac-ext-10Delta 35:       5'-...TGAAATGAGCTGATTAATCATCGGCT<UNL>GC</UNL>TATAATGTGTGGAATTG...-3'
ptac-ext-10Delta 35-7G:       5'-...TGAAATGAGCTGATTAATCATCGGCT<UNL>GC</UNL>TATAA<UNL>G</UNL>GTGTGGAATTG...-3'
ptac-ext-10Delta 35-8G:       5'-...TGAAATGAGCTGATTAATCATCGGCT<UNL>GC</UNL>TATA<UNL>G</UNL>TGTGTGGAATTG...-3'
ptac-ext-10Delta 35-9G:       5'-...TGAAATGAGCTGATTAATCATCGGCT<UNL>GC</UNL>TAT<UNL>G</UNL>ATGTGTGGAATTG...-3'
ptac-ext-10Delta 35-10G:       5'-...TGAAATGAGCTGATTAATCATCGGCT<UNL>GC</UNL>TA<UNL>G</UNL>AATGTGTGGAATTG...-3'
ptac-ext-10Delta 35-11G:       5'-...TGAAATGAGCTGATTAATCATCGGCT<UNL>GC</UNL>T<UNL>G</UNL>TAATGTGTGGAATTG...-3'
ptac-ext-10Delta 35-12G:       5'-...TGAAATGAGCTGATTAATCATCGGCT<UNL>GCG</UNL>ATAATGTGTGGAATTG...-3'
ptac-ext-10-32A:   5'-...AAATGAGCTGTT<UNL>A</UNL>ACAATTAATCATCGGCT<UNL><B>G</B>C</UNL>TATAATGTGTGGAATTG...-3'
                                         TAATGTGTGG
bub1/2: 5'-...TGAAATGAGCTGTTGACAATTAATCATCGGCTCGTA          AATTG...-3'
3'-...ACTTTACTCGACAACTCTTAATTAGTAGCCGAGCAT          TTAAC...-5'
                                         TAATGTGTGG
                                          <UNL><B>GC</B></UNL>A<UNL>C</UNL>GTGTGG
bulb1/2GC: 5'-...TGAAATGAGCTGTTGACAATTAATCATCGGCTCGTA          AATTG...-3'
3'-...ACTTTACTCGACAACTGTTAATTAGTAGCCGAGCAT          TTAAC...-5'
                                          <UNL><B>GC</B></UNL>A<UNL>C</UNL>GTGTGG
bub1/2Delta 35:       5'-...TGAAATGAGCTGATTAATCATCGGCTCGTA          AATTG...-3'
      3'-...ACTTTACTCGACTAATTAGTAGCCGAGCAT          TTAAC...-5'
                                          TAATGTGTGG
                                          ATTACACACC
bub3/4: 5'-...TGAAATGAGCTGTTGACAATTAATCATCGGCTCGTA          AATTG...-3'
3'-...ACTTTACTCGACAACTGTTAATTAGTAGCCGAGCAT          TTAAC...-5'
                                          ATTACACACC
                                          <UNL><B>GC</B></UNL>T<UNL>G</UNL>CACACC
bub3/4GC: 5'-...TGAAATGAGCTGTTGACAATTAATCATCGGCTCGTA          AATTG...-3'
3'-...ACTTTACTCGACAACTGTTAATTAGTAGCCGAGCAT          TTAAC...-5'
                                          <UNL><B>GC</B></UNL>T<UNL>G</UNL>CACACC
                                          ATTACACACC
bub3/4-11C: 5'-...TGAAATGAGCTGTTGACAATTAATCATCGGCTCGT<UNL>C</UNL>          AATTG...-3'
3'-...ACTTTACTCGACAACTGTTAATTAGTAGCCGAGCA<UNL>G</UNL>          TTAAC...-5'
                                          ATTACACACC
                                          ATTACACACC
bub3/4-12G: 5'-...TGAAATGAGCTGTTGACAATTAATCATCGGCTCG<UNL><B>G</B></UNL>A          AATTG...-3'
3'-...ACTTTACTCGACAACTGTTAATTAGTAGCCGAGC<UNL><B>C</B></UNL>T          TTAAC...-5'
                                          ATTACACACC
                                          ATTACACACC
bub3/4Delta 35:       5'-...TGAAATGAGCTGATTAATCATCGGCTCGTA          AATTG...-3'
      3'-...ACTTTACTCGACTAATTAGTAGCCGAGCAT          TTAAC...-5'
                                          ATTACACACC
                                          ATTACACACC
bub3/4-11CDelta 35:       5'-...TGAAATGAGCTGATTAATCATCGGCTCGT<UNL>C</UNL>          AATTG...-3'
      3'-...ACTTTACTCGACTAATTAGTAGCCGAGCA<UNL>G</UNL>          TTAAC...-5'
                                          ATTACACACC
                                          ATTACACACC
bub3/4-12GDelta 35:       5'-...TGAAATGAGCTGATTAATCATCGGCTCG<UNL><B>G</B></UNL>A          AATTG...-3'
      3'-...ACTTTACTCGACTAATTAGTAGCCGAGC<UNL><B>C</B></UNL>T          TTAAC...-5'
                                          ATTACACACC

The "bubble" promoters, bub1/2, bub3/4, bub1/2GC, bub3/4GC, bub3/4-11C, and bub3/4-12G were constructed by annealing two oligonucleotides each 100 nucleotides in length. The annealing mixtures contained 50 µg of each oligonucleotide in 20 mM Tris-HCl, pH 7.5, 2 mM EDTA, and 200 mM NaCl. The mixture was heated to 90 °C in a heat block and allowed to cool slowly to room temperature. The efficiency of annealing was assessed on a 3% Metaphor agarose (FMC) gel.

All of the promoter variants with a deletion of the -35 element, as well as the variants of ptac-ext-10 with point mutations were constructed using PCR for directed mutagenesis (41) with the following modifications. Primer III was complementary to primer IV. In step 1, Primers I plus IV and Primers II plus III were used to produce a segment of the final DNA product using the next previous generation of promoter variant as the template. The products of the step 1 PCR (protocol as described in the following section) were purified using the Qiaquick PCR purification kit (QIAGEN). A mixture of 5 µl of each of these products was used in a subsequent PCR step (step 2) with no additional template for two cycles. Primers I and II were then added and PCR continued for 35 more cycles. The products were purified either with the Qiaquick PCR purification kit or, in some cases, if desired band was agarose gel-isolated, purified with the Qiaquick gel extraction kit.

DNA for Nitrocellulose Filter Binding

Oligonucleotides for PCR amplification of the ptac derivatives were complementary to sequences upstream and downstream of ptac such that the promoter was approximately centered within a 100-base pair length of DNA. One of the oligonucleotides was 5'-end-labeled with 32P as already described and used to generate radioactive specific DNA (27). Labeled ptac and unlabeled competitor DNA were synthesized by PCR using 10-15 ng of template (replicative form of M13mp19ptac derivatives), 50 pmol of each primer, 10 × Taq buffer A from Fisher Scientific, and 2.5 units of either Ampli-Taq (Perkin Elmer) or Taq (Fisher) DNA polymerase in a 100-µl reaction. A Perkin Elmer Thermocycler was programmed for 35 cycles employing 94 °C for denaturation, 55 °C for annealing, and 72 °C for extension with a time of 1 min for each segment. The product was purified using the Qiaquick PCR DNA purification kit (QIAGEN), and visualized on a 3% Metaphor agarose (FMC) gel.

Nitrocellulose Filter Binding Assays

A fusion protein between regions 2-4 of sigma 70 and GST, GSTsigma (360), was used to conduct all filter binding experiments. The construction and purification of this protein has been described in detail elsewhere (27). Equilibrium competition binding assays were performed to measure the reduction in filter retention of [32P]ptac/GSTsigma (360) complexes as a function of increasing unlabeled competitor DNA (mutated versions of ptac). Binding mixtures contained 100 pM [32P]ptac and various amounts of unlabeled competitor DNA in binding buffer (10 mM Tris-HCl, pH 7.5, 14 mM potassium acetate, 0.1 mM EDTA, 1 mM dithiothreitol, 0.03% Triton X-100, and 100 µg/ml bovine serum albumin). GSTsigma (360) was added to 1 nM, followed by incubation at 37 °C for 20 min. The mixture was filtered through 24-mm nitrocellulose filters using a Hoefer model FH224V filter manifold, followed by one wash with wash buffer (10 mM Tris-HCl, 0.1 mM EDTA, and 1 mM dithiothreitol). The filters were dried and subjected to liquid scintillation counting using Ultima gold (Packard). Background DNA retention was typically 5-10% of the total and was subtracted from the values obtained for retention of complexes. All data represent the average of 2-10 duplicate experiments. On any particular day, and for any set of competitors, ptac and Delta P were always included as control competitors. Additionally, the same experiments were repeated on different days with different preparations of both DNA and protein, for every competitor. Thus, the average values presented account for the variability inherent in the filter binding measurements. As a result, the ptac and Delta P curves are the same in each figure. A more detailed discussion of this assay has been published previously (27).


RESULTS

sigma 70 Requires the First Three Positions of the -10 Element for Binding

Genetic analysis of sigma  function as part of the holoenzyme complex demonstrated that the first two positions of the -10 hexamer (TATAAT) were important specifically for promoter recognition by the sigma  subunit of RNA polymerase (29-33). The remaining four positions are clearly important for holoenzyme function, since mutations at any of these sites can severely affect promoter activity (8, 9). These positions may be important for aspects of initiation beyond promoter recognition. For example, cross-linking studies have indicated that the beta  and beta ' subunits are associated with DNA in the transcription complex (42-44). In this study, the interactions between sigma  factor and each position of the -10 hexamer were examined in the absence of the core subunits of polymerase.

We previously used equilibrium competition binding assays to show that partial polypeptides of sigma 70, lacking the N-terminal inhibitory domain, interact preferentially with promoter DNA (27, 45). In those experiments a 100-base pair DNA fragment carrying the tac promoter was mixed with increased amounts of unlabeled competitor DNA, and a fixed amount of sigma  polypeptide. Retention of sigma ·DNA complexes on nitrocellulose filters served as a measure of DNA binding. A reduction in binding to the labeled DNA by one-half required addition of either an equivalent amount of unlabeled promoter DNA or a 5-8-fold excess of unlabeled non-promoter DNA. Since the actual size of the binding site is unknown, an accurate determination of selectivity on a per-site basis cannot be ascertained. Thus, selectivity is at least 5-fold and could be as high as several hundredfold (27).

Mutations were generated at each of the six positions from -12 to -7 (TATAAT) within the -10 hexamer of the tac promoter (46) to investigate their relative importance (Table I). Equivalent substitutions have been shown previously to result in severe promoter down phenotypes in vivo (8, 9). The ability of the -10 mutant promoters to compete with the unaltered wild type tac promoter for binding to a partial sigma 70 polypeptide carrying both DNA binding domains fused to glutathione S-transferase, GSTsigma (360), was tested in vitro. Fig. 2 shows that mutations at -7, -8, or -9 had no effect on the ability of the unlabeled competitor DNA to compete with [32P]ptac for binding to GSTsigma (360). These mutated promoters competed as well as unlabeled ptac (a specific competitor). Conversely, mutations at positions -10, -11, and -12 compromised the ability of competitor DNA with these changes to compete with ptac. Competition by the -10G, -11C, and -12C or -12G DNAs more closely resembled the non-promoter competitor (Delta P). These results imply that only the upstream half of the -10 element is required for recognition and binding by the DNA binding domains of sigma  factor alone.


Fig. 2. Regions 2-4 of sigma 70 recognize positions -10, -11, and -12 of the tac promoter. End-labeled ptac DNA at 100 pM was mixed with increasing amounts of unlabeled competitor DNA. The specific competitor was ptac, the nonspecific competitor was Delta P, and point mutations in positions -7 to -12 of ptac are as indicated. Fractional retention of labeled promoter DNA as a function of increasing ratio of unlabeled competitor to labeled promoter DNA is plotted. The data shown represent the average of at least two independent experiments. Error was typically ± 15%.
[View Larger Version of this Image (23K GIF file)]


In previous analyses of competition binding, conditions were established such that at a 1:1 molar ratio of competitor DNA to ptac, filter binding was reduced by half (27). The partial sigma  factor polypeptide utilized here, GSTsigma (360), has a slightly different composition than those used for competition binding in the past. It contains all of region 2, in addition to regions 3 and 4, rather than only part of region 2 (GSTsigma (420)). This construction permits an evaluation of binding with the entire conserved region 2 present, which more closely resembles the actual DNA binding components present in wild type sigma 70. For GSTsigma (360), conditions that resulted in a 50% reduction in binding at a 1:1 ratio were not found. The basis for this observation is not clear; however, GSTsigma (360) does exhibit the expected selectivity for promoter over non-promoter DNA. The relevant comparison here is not the absolute amount of DNA needed to compete, but the relative ability of each mutant promoter DNA to compete compared to promoter (ptac) and non-promoter (Delta P) DNA.

An Extended -10 Promoter Alleviates the Requirement for a -35 Element

Certain prokaryotic promoters can function in the absence of a -35 element as long as they have the sequence TGnTATAAT, where the addition of TG creates an extended -10 promoter (10-15). Kumar et al. (47) demonstrated that removal of region 4.2 from sigma 70 eliminates the ability of holoenzyme to transcribe from a promoter with typical -35 and -10 elements, but does not destroy ability to transcribe from an extended -10 promoter. These results suggested that perhaps the extended -10 sequence can compensate for the lack of a -35 element with regard to sigma  factor binding as well.

Several variants of ptac containing an extended -10 sequence were created. ptac-ext-10 carries a substitution of G for C at -14 and a substitution of C for G at -13 to generate a "perfect" extended -10 sequence (see Table I). This promoter already contained a T at -15. A point mutation from G to A at -32 was added to the extended -10 to make ptac-ext-10-32A. This construction combines a promoter down change in the -35 hexamer, which is known to negatively affect sigma  binding (27), with the extended -10 promoter. Deletion of the -35 hexamer from ptac-ext-10 was used to generate ptac-ext-10Delta 35. These constructions allow an assessment of the contribution of the -35 element to GSTsigma (360) binding to an extended -10 promoter. Equilibrium competition binding experiments showed that, in general, as long as the promoter contained the extended -10 sequence it competed as well as ptac for binding to GSTsigma (360) (Fig. 3A). The presence of a down substitution in the -35 hexamer (ptac-ext-10-32A) did not affect binding to the extended -10 promoter but did affect binding to ptac. Thus, sigma  factor can recognize and bind to an extended -10 sequence in the absence of a -35 element, and a point mutation at -32 that normally disrupts binding by sigma  factor is ineffective in the context of an extended -10 sequence. This suggests that sigma  makes additional contacts in or near the -10 region, presumably the TG of the extended -10, that can substitute for loss of the contacts at -35.


Fig. 3. Regions 2-4 of sigma 70 bind to extended -10 promoters. End-labeled ptac DNA at 100 pM was mixed with increasing amounts of unlabeled competitor DNA. The specific competitor was ptac, and the nonspecific competitor was Delta P. Fractional retention of labeled promoter DNA as a function of increasing ratio of unlabeled competitor to labeled promoter DNA is plotted. The data shown represent the average of at least two independent experiments. Error was typically ± 15%. A, extended -10 promoter and variations with deletion of the -35 hexamer (ext-10Delta 35) and/or a point mutation from G to A at -32 (ext-10-32A), or a point mutation at -7 from T to G in addition to a deletion of the -35 hexamer (ext-10Delta 35-7G). B, extended -10 promoters with the -35 hexamer deleted in combination with point mutations in the -10 hexamer.
[View Larger Version of this Image (19K GIF file)]


Finally, the question of whether point mutations in the -10 hexamer can be tolerated in the context of an extended -10 sequence and a deletion of the -35 hexamer was examined. Point mutations were generated in the ext-10Delta 35 promoter to change each position of the -10 hexamer to a G:C base pair (see Table I). Fig. 3B shows that changes at positions -8, -9, -11, and -12 do not affect binding to an extended -10 promoter, even in the absence of a -35 element. Point mutations at -10 and -7 perhaps show a slightly deleterious effect on binding in this context, but whether this effect is significant is unclear. An evaluation of transcription initiation by holoenzyme at these promoters may be required to resolve this point.

In summary, an extended -10 promoter (TGnTATAAT) provides a very strong recognition signal for the sigma  subunit that can overcome both the lack of a -35 element as well as point mutations in the -10 element, including mutations at -11 and -12 that seriously affect binding to ptac without an extended -10 motif. The G at -14 in conjunction with the T at -15 appears to provide a highly effective sigma  factor binding determinant that allows bypass of the typical -35 and -10 determinants. Since ptac normally has a T at -15, the G at -14 in the extended -10 version may be the key additional contact point for sigma  factor. A double mutation of the G at -14 with another promoter down change may be required to eliminate binding to the extended -10 sequence.

sigma Factor Binds to ptac DNA with a Mismatch Bubble at the Transcription Start Site

Open complexes between RNA polymerase and DNA are characterized by an area of strand separation between approximately -10 and +2. Cross-linking studies have suggested that the sigma  subunit is in close proximity to one of the unpaired strands during initiation (20, 21). One possible explanation for the ability of sigma  factor to bind to DNA with mutations at -7, -8, and -9 is that these positions are only important once the strands have melted to form an open complex. To test this idea, two bubble promoter constructions were generated for use as competitor DNA. bub1/2 is identical to ptac from the upstream 5' end to position -11, and is completely base-paired along this segment. From -10 to +2, however, the two strands are non-complementary, with both corresponding to the non-template strand (see Table I). At position +3, base pairing is resumed toward the downstream end of the fragment. This creates a bubble in the promoter that mimics the transcription bubble formed by RNA polymerase holoenzyme during transcription initiation. bub3/4 is similar to bub1/2, except the unpaired segment contains the sequence corresponding to the template strand of ptac. Similar bubbles have been used by others for transcriptional analysis of RNA polymerase (48, 49).

Equilibrium competition experiments with GSTsigma (360) using mismatch bubble DNAs as competitors demonstrated that sigma  factor binds as well to bubble DNA as to duplex ptac DNA (Fig. 4A). DNA with the template (bub3/4) or non-template (bub1/2) strand in the bubble was bound similarly. These results agree with the idea that positions -11 and -12 may be the most critical sites for sigma  factor recognition of base-paired promoter DNA, as predicted by earlier genetic suppression studies. To determine if GSTsigma (360) might be binding to the bubble constructs merely because the unpaired sequence in both bub1/2 and bub3/4 is AT-rich, three changes were made at -10, -9, and -7 of both competitors to generate bubbles with higher GC content (see Table I). These DNAs, bub1/2GC and bub3/4GC, also competed well with ptac for binding (Fig. 4A). Thus it appears that once the DNA has been strand-separated to form a bubble, the sequence within the bubble is not a parameter in determining the ability of sigma  factor to bind, and unlike duplex DNA, substitution at the -10 position is tolerated.


Fig. 4. Regions 2-4 of sigma 70 bind to mismatch bubble DNA. End-labeled ptac DNA at 100 pM was mixed with increasing amounts of unlabeled competitor DNA. The specific competitor was ptac, and the nonspecific competitor was Delta P. Fractional retention of labeled promoter DNA as a function of increasing ratio of unlabeled competitor to labeled promoter DNA is plotted. The data shown represent the average of at least two independent experiments. Error was typically ± 15%. A, bub1/2 contains the mismatch with non-template sequence in the bubble, while bub3/4 contains template sequence in the bubble; bub1/2GC and bub3/4GC have three mutations that increase the GC content of the bubble (see Table I). B, bub1/2Delta 35 and bub3/4Delta 35 contain deletions of the -35 hexamer, bub3/4-11C and bub3/4-11CDelta 35 are variants with an A to C change at -11, and bub3/4-12G and bub3/4-12GDelta 35 are variants with a T to G change at -12.
[View Larger Version of this Image (20K GIF file)]


To further examine the question of whether the base-paired -11 and -12 positions in the -10 hexamer are important in the context of a bubble, point mutations at -11 from A to C and at -12 from T to G were combined with bub3/4. These changes are known to weaken promoter strength and disrupt binding by regions 2-4 to duplex promoter DNA. Surprisingly, both bub3/4-11C and bub3/4-12G competed as well as ptac for binding (Fig. 4B), indicating that in the context of a bubble, the -11 and -12 positions are not crucial for binding by sigma  factor, even though these mutations have been shown to disrupt binding to fully double-stranded DNA (27).

If none of the -10 hexamer positions are important for GSTsigma (360) to bind to a bubble promoter, then one explanation is that the -35 element provides an "anchoring" point that allows sigma  factor to bind selectively. The role of the -35 element, in the context of a bubble was addressed by constructing several additional promoter variants. The -35 hexamers of bub1/2, bub3/4, bub3/4-11C, and bub3/4-12G were deleted to generate bub1/2Delta 35, bub3/4Delta 35, bub3/4-11CDelta 35, and bub3/4-12GDelta 35. Deletion of the -35 element in the context of either bubble (bub1/2Delta 35 or bub3/4Delta 35) did not affect binding (Fig. 4B). Thus, the presence of the bubble can compensate for a lack of a -35 sequence. Interestingly, bub3/4-12GDelta 35 still competed for binding to GSTsigma (360), while bub3/4-11CDelta 35 did not. Thus, the -11 position cannot tolerate substitution in combination with a bubble and the removal of the -35 element. On the other hand, the identity of the -12 base pair in this context is flexible. In summary, GSTsigma (360) binds well to any of the bubble constructs with the only requirement being a favorable -11 base pair, providing the -35 element is absent.

sigma Factor Does Not Bind to Fully Single-stranded DNA

Potential interactions between sigma  factor and single-stranded DNA were probed using the bub1, bub2, bub3, and bub4 single-stranded oligonucleotides as competitors for GSTsigma (360) binding to double-stranded ptac. As shown in Fig. 5, none of the single-stranded DNAs was able to compete for binding with ptac. Irrespective of the sequence in the -10 region, each of these DNAs behaved like the non-promoter competitor, Delta P. Thus, the sigma  subunit prefers to interact with DNA that is either completely base-paired or partially unpaired to resemble a transcription bubble.


Fig. 5. Regions 2-4 of sigma 70 do not bind single-stranded DNA. End-labeled ptac DNA at 100 pM was mixed with increasing amounts of unlabeled competitor DNA. The specific competitor was ptac, and the nonspecific competitor was Delta P. Single-stranded competitor DNA corresponding to one strand of the tac promoter as used to construct the bubble DNA (see Table I) is indicated. Fractional retention of labeled promoter DNA as a function of increasing ratio of unlabeled competitor to labeled promoter DNA is plotted. The data shown represent the average of at least two independent experiments. Error was typically ± 15%.
[View Larger Version of this Image (20K GIF file)]



DISCUSSION

The role of the sigma  subunit of RNA polymerase in directing promoter recognition is well established. Mutations at any one of the six positions of either the -10 or the -35 promoter element can dramatically affect promoter utilization by holoenzyme (8, 9). The sigma  subunit is largely responsible for direct interactions at both the -10 and -35 sequences. Like holoenzyme, a partial sigma  polypeptide carrying both DNA binding domains is sensitive to point mutations in either the -35 or -10 hexamer, despite the fact that each DNA binding domain functions independently when present on separate polypeptides (27). It has therefore been proposed that each DNA binding domain must contact the appropriate DNA site concomitantly to form a stable protein-DNA complex (45).

Genetic analysis of sigma  function previously resulted in the identification of several mutations in region 2.4 that suppressed promoter down changes at the first two positions of the -10 hexamer (29-33). No mutations in sigma  were isolated that suppressed promoter down changes in the last four positions (-10 to -7 for ptac). Since partial polypeptides of sigma 70, in the absence of the core subunits, can be used to evaluate binding specificity (27), it became possible to directly assess promoter recognition by sigma  at the -10 element. The results presented here address the sequence requirements in the -10 region for sigma binding, the contribution of extended -10 sequences to sigma  binding, and the potential interactions between sigma  and unpaired bases in the -10 hexamer.

Base pair changes away from consensus at positions -10, -11, and -12 of ptac (TATAAT) reduce the ability of a polypeptide carrying regions 2-4 of sigma 70, GSTsigma (360), to bind to these promoters in vitro. This is consistent with the reduced transcriptional activity of holoenzyme at similarly mutated promoters in vivo. In contrast, mutations at positions -9, -8, and -7, which also compromise holoenzyme activity in vivo (9), do not affect the ability of GSTsigma (360) to bind in vitro. Similar binding selectivity was observed for a polypeptide containing region 2 only of sigma 70 (data not shown), and thus it is unlikely that the GST moiety is affecting the behavior of the sigma  portion of the fusion polypeptide.

A variant of the -10 consensus has also been characterized for GSTsigma (360) binding in which TGn appears immediately upstream of the first position of the hexamer. This extended -10 sequence can alleviate the need for a -35 consensus (13, 15, 47) or can strengthen promoters that already contain a -35 element (14, 50, 51). One explanation for the apparently improved interactions between an extended -10 promoter and RNA polymerase is that perhaps the sigma  subunit makes additional contacts with the DNA. This is supported by the observation that a C-terminal truncation of sigma 70 that removes region 4.2, the -35 recognition domain, maintains its ability to facilitate transcription initiation at an extended -10 promoter but not a standard promoter (47).

In the experiments presented here, GSTsigma (360) binds equally well to a normal ptac promoter, ptac lacking a -35 element but containing an extended -10 sequence (ptac-ext-10Delta 35), or ptac with a down mutation in the -35 element combined with an extended -10 sequence (ptac-ext-10-32A). Although these observations are consistent with the behavior of holoenzyme at similar promoters, this is the first demonstration that the sigma  subunit is the primary determinant for this mode of recognition. The binding behavior of GSTsigma (360) at ptac-ext-10Delta 35 variants with single substitutions in the -10 hexamer is not as straightforward. None of the point mutants creates a promoter that GSTsigma (360) fails to recognize to some extent, since binding is better in all cases than to non-promoter DNA (Delta P). Overall, sigma  factor may be involved in particularly strong interactions at the TG upstream of the -10 in the extended -10 promoter, and needs neither a -35 nor a perfect -10 hexamer to bind. Since the original ptac promoter has a T at -15, the G added at -14, to make the extended -10, appears to be the critical site for additional contacts between sigma  and the extended -10 sequence. These additional contacts alleviate the requirement for contacts at the -35 element, which are usually required for binding to ptac. These results imply that the extended -10 promoter is a more powerful element in directing transcription than has previously been realized.

All of the positions of the -10 hexamer are important for optimal promoter usage by holoenzyme (9, 52). Combined with the fact that positions -9 to -7 appear to be unimportant for GSTsigma (360) binding on double-stranded DNA, one possibility is that sigma  actually interacts with single-stranded DNA within the transcription bubble. In the context of holoenzyme, these positions may not be important for initial binding of polymerase to the promoter, but instead may be required for isomerization or strand separation to form a stable open complex during transcription initiation. Several studies have implicated region 2.3 of sigma  in the process of strand melting, but the precise mechanism of sigma  involvement remains obscure (34, 35, 53).

Using DNA with a mismatch bubble extending from -10 to +3, the results presented here show that sigma  can bind to these promoters when the template strand, the non-template strand, or a GC-rich version of either the template or non-template strand constitutes the sequence in the bubble. Thus, GSTsigma (360) does not seem to discriminate between base pairs in the strand-separated promoter region, including the -10 position, which is important for binding to duplex DNA. This is in contrast to recent studies indicating that holoenzyme recognizes unpaired bases in a lambda PR' heteroduplex promoter at -10 of the non-template strand (22). Additionally, a proteolytic fragment of sigma 70 containing amino acids 114-448 (C-terminal half of region 1.2 through most of region 2) forms a complex with the core subunits of polymerase that then binds to single-stranded DNA, which is comprised of the -10 sequence of the non-template strand.4 However, this sigma  fragment does not bind to DNA in the absence of core, leading to a proposal that conformational changes in the fragment upon core binding allow regions 2.3 and 2.4 to interact with the non-template single-stranded DNA in the open complex. Thus for GSTsigma (360), it is also very likely that in the absence of core, the conformation of the polypeptide does not favor specific interactions with single-stranded DNA in the -10 hexamer. The sigma  subunit appears to need to be part of a complete initiation complex in order to recognize the -10 base pairs on the non-template strand in the natural transcription bubble.

Unexpectedly, GSTsigma (360) still binds to a mismatch bubble with the -11C or -12G mutation, whereas the same change alleviates binding of double-stranded DNA. This predicts that sigma  may have a high nonspecific affinity for any single-stranded DNA. However, GSTsigma (360) does not bind to completely single-stranded DNA, regardless of the identity of the strand or the sequence in the -10 region. Finally, the importance of the -35 hexamer in the context of a mismatch bubble was evaluated. Surprisingly, deletion of the -35 elements of either bub1/2 or bub3/4 has no effect on binding by GSTsigma (360). The bub3/4Delta 35 promoter can even tolerate a promoter down point mutation at -12 and still compete for binding. However, a -11 promoter down change causes the bub3/4Delta 35 DNA to become a poor substrate for binding. Since there was no evidence for the sequence within the bubble affecting binding, bub1/2Delta 35 variants with -11 and -12 substitutions were not tested.

The sigma  subunit appears to recognize the -10 hexamer in two different contexts, in the presence of the core subunits as holoenzyme, and in the absence of core. The results presented here are consistent with the following scenario for recognition of promoter DNA by sigma  factor. First, holoenzyme is sensitive to changes at all six positions of the -10 and also requires a -35 element to firmly "anchor" the protein to the promoter. At an extended -10 promoter, holoenzyme makes extra contacts with the -10 region of the DNA, eliminating the need for a -35 anchor. sigma  factor, in the absence of core, appears to be involved in interactions with the first three positions of the -10 element (TATAAT) in binding of double-stranded DNA. The identity and single or double-stranded character of the remaining three positions (-7 to -9) is not a factor in recognition by sigma  as long as the -35 element is present. Second, the presence of an extended -10 very effectively replaces the requirement for a -35 element presumably by allowing sigma  to make additional promoter contacts that stabilize the complex. While the substitution of an extended -10 for a poor or absent -35 has been invoked in the past to explain promoter activity, this is the first demonstration that the sigma  subunit is primarily responsible for utilization of the extended -10 motif.

Finally, while sigma  does not bind to single-stranded DNA, it has an expectedly high affinity for DNA that is primarily double-stranded but that contains a mismatch bubble in the -10 sequence. The specificity of sigma  for specific bases in the single-stranded transcription bubble that was demonstrated by Roberts and Roberts (22) was not observed using a partial polypeptide of sigma  alone, in agreement with the behavior of a proteolytic fragment of sigma .4 It appears likely that interactions between sigma  and core dictate this mode of recognition. During transcription initiation, sigma  participates in steps beyond initial promoter binding. It has recently become clear the Regions 1.1 and 1.2 are involved in transitions following promoter binding.1 Thus, the conformation and positioning of sigma  on the DNA may change at each step of initiation. Since the behavior of holoenzyme at the various promoters utilized here, for either DNA binding or transcription initiation, has not been analyzed, the contribution of the core subunits to utilization of these promoters remains to be determined.

The recognition of promoters by RNA polymerase is a complex subject. There are, of course, promoters that contain a weak match to the -10 consensus yet function in transcription. However, many parameters other than the sequence of the hexamers come into play in determining the overall efficiency of a particular promoter (54). These include the DNA topology in the region, participation of activators or other DNA-binding proteins, and interaction of the alpha  subunit of polymerase with DNA upstream of the promoter. Interestingly, recent studies of the promoter recognition properties of the stationary phase sigma , sigma 38 or sigma S, indicate that the -10 hexamer but not the -35 is essential for transcription initiation from sigma S-dependent promoters (55-57). It is clear that the -10 hexamer plays an important role in specifying transcriptional start sites and that the sigma  subunit of RNA polymerase has a primary role in utilization of specific sequences at that region of the promoter.


FOOTNOTES

*   This work was supported in part by Grant 5RO1AI19635-14 from the National Institutes of Health (to Carol A. Gross, University of California, San Francisco), Grant 94G-263 from the Texas Affiliate of the American Heart Association (to A. J. D.), and Grant NP-902 from the American Cancer Society (to A. J. D.). The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Dagger    To whom correspondence should be addressed: Dept. of Microbiology and Molecular Genetics, University of Texas Health Science Center, 6431 Fannin JFB1.765, Houston, TX 77030. Tel.: 713-794-1744 (ext. 1511); Fax: 713-794-1782; E-mail: dombros{at}utmmg.med.uth.tmc.edu.
1    Wilson, C., and Dombroski, A. J. (1997) J. Mol. Biol., in press.
2    Sequences are available from the author upon request.
3    The abbreviations used are: PCR, polymerase chain reaction; GST, glutathione S-transferase.
4    Severinova, E., Severinov, K., Fenyo, D., Marr, M., Brody, E. N., Roberts, J. W., Chait, B. T., and Darst, S. A. (1997) J. Mol. Biol. 263, 637-647.

Acknowledgments

I thank Dr. C. Gross, in whose laboratory this work began, for advice, encouragement, and critical reading of the manuscript; Dr. P. deHaseth for suggestions for making mismatch bubble DNA; and the members of my laboratory for helpful discussions.


REFERENCES

  1. Burgess, R. R., Travers, A. A., Dunn, J. J., and Bautz, E. K. F. (1969) Nature 221, 43-44 [Medline] [Order article via Infotrieve]
  2. Burgess, R. R., and Travers, A. A. (1970) Fed. Proc. 29, 1164-1169 [Medline] [Order article via Infotrieve]
  3. Saucier, J., and Wang, J. (1972) Nat. New Biol. 239, 167-170 [Medline] [Order article via Infotrieve]
  4. Gross, C. A., Lonetto, M., and Losick, R. (1992) Transcriptional Regulation, pp. 129-176, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY
  5. Lonetto, M., Gribskov, M., and Gross, C. (1992) J. Bacteriol. 174, 3843-3849 [Medline] [Order article via Infotrieve]
  6. Hawley, D. K., and McClure, W. R. (1980) Proc. Natl. Acad. Sci. U. S. A. 77, 6381-6385 [Abstract]
  7. Harley, C. B., and Reynolds, R. P. (1987) Nucleic Acids Res. 15, 2343-2361 [Abstract]
  8. Youderian, P., Bouvier, S., and Susskind, M. M. (1982) Cell 30, 843-853 [Medline] [Order article via Infotrieve]
  9. Moyle, H., Waldburger, C., and Susskind, M. M. (1991) J. Bacteriol. 173, 1944-1950 [Medline] [Order article via Infotrieve]
  10. Backendorf, C., Brandsma, J. A., Kaitasova, T., and van dePutte, P. (1983) Nucleic Acids Res. 11, 5795-5810 [Abstract]
  11. Inokuchi, K., Furukawa, H., Nakamura, K., and Mizushima, S. (1984) J. Mol. Biol. 178, 653-668 [Medline] [Order article via Infotrieve]
  12. Ponnambalam, S., Webster, C., Bingham, A., and Busby, S. (1986) J. Biol. Chem. 261, 16043-16048 [Abstract/Free Full Text]
  13. Keilty, S., and Rosenberg, M. (1987) J. Biol. Chem. 262, 6389-6395 [Abstract/Free Full Text]
  14. Chan, B., and Busby, S. (1989) Gene (Amst.) 84, 227-236 [CrossRef][Medline] [Order article via Infotrieve]
  15. Xiong, X. F., and Reznikoff, W. S. (1993) J. Mol. Biol. 231, 569-580 [CrossRef][Medline] [Order article via Infotrieve]
  16. Siebenlist, U. (1979) Nature 279, 651-652 [Medline] [Order article via Infotrieve]
  17. Siebenlist, U., Simpson, R. B., and Gilbert, W. (1980) Cell 20, 269-281 [Medline] [Order article via Infotrieve]
  18. Kirkegaard, K., Buc, H., Spassky, A., and Wang, J. C. (1983) Proc. Natl. Acad. Sci. U. S. A. 80, 2544-2548 [Abstract]
  19. Sasse-Dwight, S., and Gralla, J. D. (1989) J. Biol. Chem. 264, 8074-8081 [Abstract/Free Full Text]
  20. Simpson, R. B. (1979) Cell 18, 277-285 [Medline] [Order article via Infotrieve]
  21. Buckle, M., and Buc, H. (1994) Transcription: Mechanisms and Regulation, Raven Press, Ltd., New York
  22. Roberts, C. W., and Roberts, J. W. (1996) Cell 86, 495-501 [Medline] [Order article via Infotrieve]
  23. DeHaseth, P. L., and Helmann, J. D. (1995) Mol. Microbiol. 16, 817-824 [Medline] [Order article via Infotrieve]
  24. Stragier, P., Parsot, C., and Bouvier, J. (1985) FEBS Lett. 187, 11-15 [CrossRef][Medline] [Order article via Infotrieve]
  25. Gribskov, M., and Burgess, R. R. (1986) Nucleic Acids Res. 14, 6745-6763 [Abstract]
  26. Helmann, J. D., and Chamberlin, M. J. (1988) Annu. Rev. Biochem. 57, 839-872 [CrossRef][Medline] [Order article via Infotrieve]
  27. Dombroski, A. J., Walter, W. A., and Gross, C. (1992) Cell 70, 501-512 [Medline] [Order article via Infotrieve]
  28. Dombroski, A. J., Walter, W. A., and Gross, C. A. (1993) Genes Dev. 7, 2446-2455 [Abstract]
  29. Siegele, D. A., Hu, J. C., Walter, W. A., and Gross, C. A. (1989) J. Mol. Biol. 206, 591-603 [Medline] [Order article via Infotrieve]
  30. Zuber, P., Healy, J., Carter, H. L., III, Cutting, S., Moran, C. P., Jr., and Losick, R. (1989) J. Mol. Biol. 206, 605-614 [Medline] [Order article via Infotrieve]
  31. Daniels, D., Zuber, R., and Losick, R. (1990) Proc. Natl. Acad. Sci. U. S. A. 87, 8075-8079 [Abstract/Free Full Text]
  32. Waldburger, C., Gardella, T., Wong, R., and Susskind, M. M. (1990) J. Mol. Biol. 215, 267-276 [Medline] [Order article via Infotrieve]
  33. Kenney, T. J., and Moran, C. P., Jr. (1991) J. Bacteriol. 173, 3282-3290 [Medline] [Order article via Infotrieve]
  34. Juang, Y.-L., and Helmann, J. D. (1994) J. Mol. Biol. 235, 1470-1488 [CrossRef][Medline] [Order article via Infotrieve]
  35. Rong, J. C., and Helmann, J. D. (1994) J. Bacteriol. 176, 5218-5224 [Abstract]
  36. Gardella, T., Moyle, H., and Susskind, M. M. (1989) J. Mol. Biol. 206, 579-590 [Medline] [Order article via Infotrieve]
  37. Makino, K., Amemura, M., Kim, S.-K., Nakata, A., and Shinagawa, H. (1993) Genes Dev. 7, 149-160 [Abstract]
  38. Li, M., Moyle, H., and Susskind, M. M. (1994) Science 263, 75-77 [Medline] [Order article via Infotrieve]
  39. Kuldell, N., and Hochschild, A. (1994) J. Bacteriol. 176, 2991-2998 [Abstract]
  40. Kunkel, T. A., Roberts, J. D., and Zakour, R. A. (1987) Methods Enzymol. 154, 367-382 [Medline] [Order article via Infotrieve]
  41. Ito, W., Ishiguro, H., and Kurosawa, Y. (1991) Gene (Amst.) 102, 67-70 [CrossRef][Medline] [Order article via Infotrieve]
  42. Harrison, C. A., Turner, D. H., and Hinkle, D. C. (1982) Nucleic Acids Res. 10, 2399-2414 [Abstract]
  43. Hillel, Z., and Wu, C.-W. (1978) Biochemistry 17, 2954-2961 [Medline] [Order article via Infotrieve]
  44. Brodolin, K. L., Studitsky, V. M., and Mirzabekov, A. D. (1993) Nucleic Acids Res. 21, 5748-5753 [Abstract]
  45. Dombroski, A. J., Johnson, B. D., Lonetto, M., and Gross, C. A. (1996) Proc. Natl. Acad. Sci. U. S. A. 93, 8858-8862 [Abstract/Free Full Text]
  46. Amann, E., Brosius, J., and Ptashne, M. (1983) Gene (Amst.) 25, 167-178 [Medline] [Order article via Infotrieve]
  47. Kumar, A., Malloch, R. A., Fujita, N., Smillie, D. A., Ishihama, A., and Hayward, R. S. (1993) J. Mol. Biol. 232, 406-418 [CrossRef][Medline] [Order article via Infotrieve]
  48. Aiyar, S. E., Helmann, J. D., and deHaseth, P. L. (1994) J. Biol. Chem. 269, 13179-13184 [Abstract/Free Full Text]
  49. Ring, B. Z., Yarnell, W. S., and Roberts, J. W. (1996) Cell 86, 485-493 [Medline] [Order article via Infotrieve]
  50. Grana, D., Youderian, P., and Susskind, M. M. (1985) Genetics 110, 1-16 [Abstract/Free Full Text]
  51. Ponnambalam, S., Chan, B., and Busby, S. (1988) Mol. Microbiol. 2, 165-172 [Medline] [Order article via Infotrieve]
  52. Danot, O., and Raibaud, O. (1994) J. Mol. Biol. 238, 643-648 [CrossRef][Medline] [Order article via Infotrieve]
  53. Jones, C. H., and Moran, C. P., Jr. (1992) Proc. Natl. Acad. Sci. U. S. A. 89, 1958-1962 [Abstract]
  54. Knaus, R., and Bujard, H. (1990) Nucleic Acids Mol. Biol. 4, 110-122
  55. Hiratsu, K., Shinagawa, H., and Makino, K. (1995) Mol. Microbiol. 18, 841-850 [Medline] [Order article via Infotrieve]
  56. Tanaka, K., Kusano, S., Fujita, N., Ishihama, A., and Takahashi, H. (1995) Nucleic Acids Res. 23, 827-834 [Abstract]
  57. Kolb, A., Kotlarz, D., Kusano, S., and Ishihama, A. (1995) Nucleic Acids Res. 23, 819-826 [Abstract]
  58. Lesley, S. A., and Burgess, R. R. (1989) Biochemistry 28, 7728-7734 [Medline] [Order article via Infotrieve]

©1997 by The American Society for Biochemistry and Molecular Biology, Inc.