T H E J O U R N A L OF BIOLOGICAL CHEMISTRY 0 1989 by The American Society for Blochemistry and Molecular Biology, InC

Vol. 264, No. 18, Lssue of June 25, pp. 10451-10456, 1989 Printed in U.S.A.

Rotational Orientation of Upstream Curved DNA Affects Promoter Function in Bacillus subtilis*

(Received for publication, September 6, 1988)

Carl F. McAllisterS and Eric C. Achberger From the Department of Microbiology, Louisiana State Uniuersity, Baton Rouge, Louisiana 70803

The Alu156 promoter isolated from the Bacillus sub- tilis bacteriophage SP82 is dependent on curved DNA upstream of the -35 region for efficient function. Short DNA insertions of 6-29 base pairs were used to simultaneously change the linear placement and rota- tional orientation of this curved DNA relative to the -35 region. When these mutant promoters were ana- Iyzed in vivo using transcriptional fusions with a chlor- amphenicol acetyltransferase gene, changes in the ro- tational orientation of the curved DNA correlated with changes in promoter function. The most efficient mu- tant promoters contained insertions of l l and 2 1 base pairs, and insertions of 15 and 25 base pairs resulted in the least efficient mutant promoters. The impor- tance of the proper rotational alignment of the curved DNA to promoter activity was also observed in vitro at the level of transcription of RNA polymerase bind- ing. Based on the electrophoretic mobilities of DNA fragments containing the various insertion mutant promoters, there was a second region of curved DNA downstream of the insertion point. The findings are consistent with the idea that the curved DNA deflects the helix back toward the promoter-bound RNA po- lymerase molecule to allow the enzyme to interact di- rectly with upstream DNA. These interactions are pro- posed to structure the DNA for the formation of the open promoter complex.

The existence of static curves in DNA is well documented (1-4). In a number of cases, curved DNA is characterized by runs of adenines with a periodicity of approximately 10-11 bp.’ The degree of curvature of these regions has been esti- mated based on altered electrophoretic mobility of DNA frag- ments (1-3) and intramolecular ligation rates (5, 6). Se- quence-dependent curvature has been observed in regions of DNA involved in site-specific recombination (7, 8), the initi- ation of DNA replication (6, 9-11), and transcription initia- tion.

DNA curvature and protein-mediated DNA bending are required elements for efficient transcription of some pro- moters. DNA bending has been implicated in transcription activation of the lactose operon promoter PI by the CAMP-

* 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.

The nucleotide seguence(s) reported in this paper has been submitted

J03904. to the GenBankTM/EMBL Data Bank with accession number($

$ Present address: Center for Infectious Diseases, CDC, 1600 Clif- ton Rd., N. E., Atlanta, GA 30033.

EDTA. ‘The abbreviations used are: bp, base pairs; TBE, Tris-borate-

CAMP receptor protein complex (12). DNA looping has been proposed as part of the activation process of the Klebsiella pneumoniae nifH promoter by the nifA gene product (13) and the Escherichia coli glnAp2 promoter by the glnG gene product (14). Curved DNA has been identified upstream of the E. coli ompF promoter in a region that overlaps the binding site of the transcription activator OmpR (15). DNA curvature ap- pears to stimulate transcription in the absence of ancillary proteins. Mutation of the curved DNA upstream from pro- moters for a number of tRNA and rRNA operons from Gram- negative bacteria (15-19) and the bacteriophage h PL pro- moter (20, 21) reduced transcription. In addition, nucleotide sequence analysis suggests a correlation between promoters displaying high rates of transcription initiation and those with specific distributions of nucleotides associated with DNA curvature (22).

Runs of adenines upstream from promoters of Gram-posi- tive bacteria, such as Bacillus subtilis, are a common feature (23, 24). Using a deletion analysis, it was shown that the A+T-rich DNA upstream of the spoVG promoter was required for efficient transcription by the u3’ and u37 forms of the 3. subtilis RNA polymerase (25). Curved DNA upstream of two B. subtilis bacteriophage SP82 promoters enhanced binding of the major RNA polymerase (i.e. the containing enzyme) and stimulated in vivo and in vitro transcription (26). In particular, one of these promoters, the Alu156 promoter, was strongly dependent on the curved DNA for efficient function. This promoter was selected for further study.

While it has been established that curved DNA can stimu- late RNA polymerase binding to promoters, the nature of these interactions is unresolved. This report described the correlation between the precise orientation of the curved DNA relative to the promoter proper and the ability of this DNA to function as a cis-acting element for enhanced transcription from the Alu156 promoter. This structure/function relation- ship was investigated using mutant promoters created by the insertion of oligonucleotides between the polyadenine-in- duced curve and the -35 region of the promoter. Each pro- moter was analyzed by measuring in vivo and in vitro tran- scription and RNA polymerase holoenzyme binding.

EXPERIMENTAL PROCEDURES’

RESULTS

The Alu156 promoter possesses polyadenine-induced DNA curvature upstream of the -35 region (Fig. 1). Previous stud- ies support the idea that the curved DNA is required for efficient utilization of this promoter in B. subtilis (26). To

Portions of this paper (including “Experimental Procedures,” Fig. 6, and Table I) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are included in the microfilm edition of the Journal that is available from Waverly Press.

10451

10452 Positioning of Upstream DNA and Promoter Selection -150 -140

t

GMTTCCCCGGATCCGTC

-130 -1tO -110 - 1 p O -?O -80 -70 8 I

I I , CTTCCMGlVUGATATCCTAACAGCACMGAGCGGAGCGG~CACG~TTCTA~TCCAGAACMCCTCTG

-60 -70 -40 -30 -20 - 10 tl I I I

I I I I CT~~CCTGAAMA~GCAAAMGTTG~Cll'TCTCTACGAGGTGTGGCATMTMTCTTAAC

t10 +20 +30 I I I

AACAGC~ACGCTA&ACGSATCC&MTCC

FIG. 1. Nucleotide sequence of the Alu156 promoter. Bases representing the +1 transcription start site, the -35 region, and the -10 region are in bold type. The runs of adenines in the nucleotide sequence immediately upstream of the -35 region are underlined.

A h 1 5 6

t 6

+9

t 1 1

t 1 3

+15

+17

+ l Q

t 2 1

+25

+29

-85 -75 -65 -55 -45 -35 , C A T C C A G M C M C C T C n ; C T M C C T G A M A A ~ ~ ~ T T G A C T

G M C M C C T C T C C T M M T T C C T " G ~ T T ~ G A C T

C M C C T C T G C T M M l T C C T ~ ~ G ~ C C G A C G A ~ A C T

A C C T C T G C T M M T T C C n ~ l T G T T C C G G A T C G A T T G A C T

C T C T G C T M M ~ C C n ~ ~ C C f f i A ~ ~ G A T T

CTGCTMAAlTCCTGAMAATTI%WTXTKCGGCU%ATCGATTGACT

G C T M M ~ C C T T T C ~ T C G T C G A C G A T T G A C T

T M M ~ C C ~ A M A A ~ " G T T G ~ C f f i A T C G T ~ C G A C G A T T G A C T

) r M T F C C T C A M A A ~ C C M C C G G A T C G T C G A T C G A C G A T T G A C T

TCCTGhMUTTI%C"GlTGTTCCGGCCGGATCGTCGATCGACGATCGACT

G h M U ~ ~ ~ C C G G A T C G ~ C G G A T C G T C G A T C G A C G A T T G A C T

Displacement Rotational

0.0

0.57

0.86

1.05

1.24

1.43

1.62

1 .81

2.0

2.38

2.76

FIG. 2. Nucleotide sequence of DNA immediately upstream from the -35 region of the Alul56-derived mutant promoters. The numbering of the nucleotide sequence corresponds to the +1 transcription start site of the Alu156 promoter presented in Fig. 1. The numerical designation of each promoter is listed in the right column. The oligonucleotide insertions for each mutant are under- lined, and the runs of adenines are in bold type. The rotational displacement based on B-form DNA is given in helical turns for each.

characterize the stimulation of transcription by curved DNA, a series of mutant promoters was constructed by the insertion of synthetic, double-stranded oligonucleotides of various lengths between the -35 region and the curved upstream DNA of the Alu156 promoter. These insertions served a dual purpose. First, they increased the distance between the curved DNA and the promoter by discrete increments, and second, they rotated the curved DNA around the helix with respect to the -35 region. The relevant sequence of these mutant promoters as well as the rotational displacement of the poly- adenine regions for each mutant are presented in Fig. 2. For each 10.5 bp of inserted DNA, there would be a linear dis- placement of 3.4 nm and a rotational displacement of 360" (i.e. one full helical turn).

Electrophoretic Mobilities of Promoter-containing DNA Fragments-It has previously been shown that DNA frag- ments containing the Alu156 promoter migrated in 8% poly- acrylamide gels at a rate consistent with a molecule 22% larger than its actual size (26). Deletion or replacement of the upstream DNA significantly reduced but did not entirely eliminate this aberrant mobility. This suggests that in addi- tion to the upstream regions, the sequence downstream from the HincII cut site at -33 also exhibited some curvature. Thus, it was hypothesized that rotation of two curves relative to one another would result in a detectable change in the structure of the promoter DNA fragment. To test for altered DNA conformations, the promoters constructed in this study were excised from the vector with EcoRI and PstI and elec- trophoresed on 10% polyacrylamide gels at three different

temperatures. It has been reported that the electrophoretic mobility of curved DNA returns to normal at elevated tem- peratures (19, 40). In agreement with this observation, a t 65 "C, DNA fragments containing the various promoters mi- grated at rates that were most consistent with those predicted from the nucleotide sequence derived lengths (Fig. 3). The electrophoretic mobilities of these DNA fragments were most severely retarded at 5 and 23 "C. The finding that electropho- retic mobilities were not completely restored to normal even at 65 "C is consistent with the existence of a highly stable, curved DNA molecule.

Interestingly, the ratio of apparent/actual length of DNA fragments does not remain constant, as would be predicted for linear molecules of increasing length. The cyclical pattern observed in Fig. 3 would be expected if there is a region of curved DNA on both sides of the rotation point. Decreasing end-to-end distance directly correlates with decreasing elec- trophoretic migration rates for DNA fragments in polyacryl- amide gels (41). When two regions of curved DNA are present on a single molecule, the electrophoretic mobility would be retarded most when the two curves are in the same orientation (Le. a U-shaped DNA molecule). The electrophoretic mobility would be closest to normal when the two curves are direction- ally opposed forming an S-shaped molecule (41). The observed cyclical pattern had a periodicity of approximately one helical turn; the ratio of apparent/actual DNA fragment length de- termined at 23 "C reached a maximum for the +6 and +17 mutant promoter DNA fragments and a minimum for the +15 and +25 mutant promoters. The +6 and +17 fragments should have the polyadenine regions on the opposite side of the helix relative to the wild type and +11 and +21 mutants.

If there were a single DNA curve on either side of the point of rotation, one would expect to see the points for each curve in Fig. 3 arranged so that a single sine curve be drawn through them. The data, however, show a shoulder on the right side of each peak which is most pronounced a t 5 "C. The most likely explanation for these patterns is the presence of a third region of curvature whose stability is enhanced at 5 "C.

I n Vivo Expression from Alu156 Promoter Constructions- Transcriptional fusions were established between each of the

g 1.01 I Q: 0 5 10 15 20 25 30

Insert Length (bp)

FIG. 3. Effect of temperature on the electrophoretic mobil- ity of DNA fragments containing the Alu156 promoter Or insertion mutant promoters. The ratio of the apparent length based on electrophoretic mobility to actual length determined from the nucleotide sequence is presented as a function of the length of the oligonucleotide insertion for the wild-type Alu156 and insertion mutant promoters. The apparent length of promoter-containing DNA fragments was determined using pBR322-MspI and pBR322-HwIII size standards as described under "Experimental Procedures." The temperatures used during electrophoresis were 5 "C (a), 23 "c (o), and 65 "C (A).

Positioning of Upstream DNA and Promoter Selection 10453

promoters and the chloramphenicol acetyltransferase gene carried on the promoter cloning vector pPL703. The levels of chloramphenicol acetyltransferase specific activity measured for each promoter construction in B. subtilis are presented in Fig. 4-4. From the data in Fig. 4A, it was apparent that efficient transcription from the Alu156 promoter was dependent, in part, on the rotational orientation of the upstream DNA relative to the promoter proper. The mutant constructions which allowed the highest levels of chloramphenicol acetyl- transferase expression were the promoters in which the up- stream DNA was rotated by one or two turns of the helix (e.g. +11 and +21 mutant promoters) and was, thus, in the same

1 1 1 0 5 10 15 2 0 2 5 30

Inser t ion Length (bp) FIG. 4. Effect of oligonucleotide insertions on the utilization

of the Alu156 promoter in B. subtilis. Panel A, promoter-de- pendent chloramphenicol acetyltransferase (CAT) specific activity as a function of the length of the insertions in mutant promoters. Chloramphenicol acetyltransferase specific activity, micromoles of chloramphenicol acetylated per min/mg protein at 25 “C, for each promoter construction was determined from three independent trans- formants containing the correct pPL703-promoter transcriptional fusions. The values were corrected for the background measurements with pPL703-containing B. subtilis 1A510. Panel B, RNA polymerase binding to promoter-containing DNA fragments expressed as the fraction of input DNA bound normalized to the binding observed for the internal control, the Alu156 promoter-containing fragment. The normalized binding determined at an RNA polymerase/DNA weight ratio of 2.5 is presented as a function of the oligonucleotide insertion length. The conditions for the competition binding assay are described under “Experimental Procedures.” Panel C, the effect of oligonucle- otide insertion length on in vitro RNA synthesis measured by the incorporation of [3H]UTP into acid-precipitable material as described under “Experimental Procedures.” Purified RNA polymerase was used to transcribe either supercoiled (A) or linear (A) pUC8 deriva- tives containing the individual promoter constructions at an enzyme/ DNA weight ratio of 2.7.

rotational orientation as the wild-type Alu156 promoter. Chloramphenicol acetyltransferase specific activities de- creased to a minimum at or very near the point at which the polyadenine regions would be on the opposite side of the DNA helix from the -35 region (i.e. the +15 and +25 insertions). The high chloramphenicol acetyltransferase specific activity observed from the +6 and +9 mutant promoters may reflect a rotation-independent effect based on the proximity of the polyadenine region to the promoter proper.

In Vitro Binding of RNA Polymerase to Insertion Mutant Promoters-To ascertain if the changes in in vivo promoter function reflected differences in the affinity of RNA polym- erase for the various promoter constructions, the binding of purified B. subtilis RNA polymerase (core-0~~6) to the wild- type and mutant promoters was examined. The wild-type promoter was allowed to compete with various combinations of three to four mutant promoters for subsaturating levels of RNA polymerase. Equal molar amounts of each promoter were used in these nitrocellulose filter binding assays. RNA polymerase binding data presented in Fig. 4B represent results from a single enzyme concentration in the competition bind- ing assay. These data have been normalized to the level of RNA polymerase binding observed with the Alu156 promoter- containing fragment which was present in each assay. An example of the competition binding assay is presented in Fig. 5. Low recovery of the weakly bound promoters was observed until the promoters with a high affinity for RNA polymerase were bound at near-maximum level. The normalized data presented in Fig. 4B were generated at an enzyme/DNA weight ratio of 2.5 (e.g. equivalent to the 15-pg point in Fig. 5 ) .

The binding data agree well with the results obtained from the in vivo expression assays. Maximum binding of RNA polymerase occurred with the wild type Alu156 promoter and the +11 and +21 mutants which retained the original rota- tional orientation of the upstream DNA relative to the -35 region. Insertions that altered this orientation reduced the affinity of RNA polymerase for the mutant promoters. In addition, a mutant promoter that lacked the upstream DNA, the Alu156 extended promoter (26), was included in the competition binding assay. For this mutant, the fraction of input DNA bound was 0.12 (normalized as in Fig. 4B) at an enzyme/DNA ratio of 2.5. This level of binding was approxi- mately one-half that observed for the least efficient insertion

0.81 I

R N A Polymerase (pg) FIG. 5. Relative affinity of B. subtiiis RNA polymerase for

DNA fragments containing the Ah156 promoter or insertion mutant promoters. Representative results from the competition binding assay measuring the fraction of promoter-containing DNA retained on a nitrocellulose membrane as a function of RNA polym- erase concentration. At each RNA polymerase concentration, the reaction contained equal amounts of DNA fragments containing the Alu156 promoter (O), the +6 mutant (O), the +11 mutant (A), and +15 mutant (A). RNA polymerase binding was quantified as described under “Experimental Procedures.”

10454 Positioning of Upstream DNA and Promoter Selection

mutant promoters, the +6 and +15 mutants. This suggests that while the rotational orientation of upstream and down- stream curved DNA was very important in producing a DNA conformation conducive to efficient RNA polymerase binding, the presence of the curved DNA in any orientation affects binding.

There were two apparent discrepancies between the in vivo data and the in vitro RNA polymerase binding data. First, the +6 mutant appeared to bind RNA polymerase weakly. If the polyadenine region can affect promoter function by its prox- imity to the promoter, this effect was not evident at the level of RNA polymerase binding to a linear DNA fragment. Sec- ond, the +13 mutant promoted very low chloramphenicol acetyltransferase expression in vivo but appeared to bind RNA polymerase tightly. In separate experiments, 400 PM ATP and CTP were included in the competition binding reactions to permit transcription initiation. Just prior to filtration, these reactions were diluted in reaction buffer containing 1 M NaCl at 0 "C to dissociate non-initiated complexes (42). Filter re- tention of DNA fragments containing the +13 mutant relative to other promoters in the reaction was unchanged from that presented in Fig. 4B (data not shown). This suggests that RNA polymerase was able to form initiatable, open promoter complexes at the +13 mutant promoter sequence as efficiently as it could to the other promoters in this study.

Effects of Insertion Mutagenesis of the Alu156 Promoter on in Vitro Transcription-The ability of the wild-type and mutant promoters to direct transcription from linear and naturally supercoiled pUC8-derivative templates was tested using the same enzyme/DNA weight ratios as in the RNA polymerase binding experiments. In general, the data from in vitro transcription studies agreed with the in vitro binding and in vivo chloramphenicol acetyltransferase expression findings. Promoters which gave the most efficient RNA syn- thesis (e.g. the Alu156 promoter and the +11 and +21 mu- tants) also bound RNA polymerase more tightly and permit- ted higher levels of in vivo chloramphenicol acetyltransferase expression. Again, the promoters which gave the lowest tran- scription rates in vitro were those in which the polyadenine regions had been rotated to the opposite side of the helix as the -35 region.

The levels of transcription from linear versus supercoiled templates correlated to the lengths of the respective tran- scripts. The RNA synthesized from the linear transcripts was approximately 900 bases long, whereas transcription from the supercoiled template would be expected to generate a tran- script of at least 1600 bases in length if terminated efficiently at the ApR gene terminators (43). The increased incorporation of UMP seen with the supercoiled templates is fairly consist- ent through the series of insertion mutants when compared t o the levels from linear templates for each construction. The notable exceptions are the +15 and +25 insertion mutants. The partial untwisting of the negatively supercoiled DNA template seems to stimulate transcription from these mutant promoters. The linear displacement of the curved DNA from the promoter may contribute to this effect because transcrip- tion was not preferentially elevated on supercoiled DNA templates for the +6 mutant and the effect was greater for the +25 mutant than for the +15 mutant (Fig. 4C).

DISCUSSION

The orientation of the curved DNA immediately upstream of the -35 region was altered by the introduction of short oligonucleotides between the curved DNA and the rest of the promoter. The insertion of these oligonucleotides resulted in both linear and rotational displacement of the upstream DNA

in relation to the promoter. In general, insertion mutant promoters in which the curved DNA was rotated away from the original orientation displayed reduced function. Insertions of 11 or 21 bp that maintained the original rotational orien- tation while linearly displacing the curved DNA produced mutant promoters that retained most of the activity of the wild-type promoter. These results suggest that proper rota- tional orientation is the most critical factor for the stimulation of transcription by the upstream curved DNA.

There was a minor component of transcription enhance- ment based on the proximity on the curve to the promoter. The insertion mutant promoter that displayed the lowest activity (ie. chloramphenicol acetyltransferase specific activ- ity, RNA polymerase binding, and in vitro transcription) retained more of the original promoter function than mutant promoters in which the upstream curved DNA was deleted (26). This indicates that the curved DNA could stimulate transcription (e.g. approximately 2-fold) even when the rota- tional orientation was not optimal. This may explain, in part, why the +6 and +9 insertion mutants exhibited efficient promoter function in vivo.

The proximity of the curved DNA to the promoter may also govern the effect of DNA supercoiling on promoter func- tion. DNA supercoiling has been shown to stimulate tran- scription from Alu156 mutant promoters lacking the curved upstream DNA (26). For the most part, this effect was not observed for insertion mutant promoters. Slightly elevated transcription was observed on supercoiled templates contain- ing the +15 mutant; this effect was more pronounced for the +25 mutation. While curved DNA has been shown to influ- ence the topology of supercoiled molecules (44), it is not known how this may affect protein-DNA interactions.

Polyacrylamide gel electrophoresis of DNA fragments con- taining the various Alu156 promoter constructions demon- strated the presence of a second region of curved DNA in addition to the polyadenine-dependent curvature just up- stream of the -35 region. The cyclical pattern of the curve generated in a plot of apparent fragment length/actual length versus insert length (Fig. 3) indicated the presence of a second curve downstream of the point of rotation (ie. the site of the insertion). This second curve may be induced by the pairs of adenines at -5 and +5 (Fig. 1). Adenine dinucleotides asso- ciated with the -10 region have been proposed as a possible site of DNA curvature (4, 45). Also, the asymmetry of the cyclical pattern presented in Fig. 3 suggests the presence of a third region of altered DNA conformation. The DNA up- stream of -90 is characterized by the presence of runs of 2-4 adenines and thymines. The length and periodicity along the helix of these runs of nucleotides marginally conforms to the sequence motif associated with curved DNA (3, 5). It would not be unexpected that the DNA curvature induced by this sequence would be stabilized at low temperature as depicted in Fig. 3. Although it is not known if DNA upstream of -80 is important for transcription from the Alu156 promoter, replacement of DNA upstream of -33 with the analogous DNA from the Ball29 promoter resulted in the loss of only 35% of the function in vivo (26). This curved upstream DNA from the Ball29 promoter terminates at -77 and lacks DNA for additional upstream curves.

Several models have been suggested to explain the role of curved DNA in the activation of transcription. It has been proposed that the curved upstream DNA acts as a sequence- independent RNA polymerase binding site to maintain a locally high concentration of enzyme and, hence, facilitate binding to the promoter (22, 46). This model cannot ade- quately account for the effect of curved DNA on the stability

Positioning of Upstream DNA and Promoter Selection 10455

of RNA polymerase-promoter complexes once formed (26) or the effect of altered rotational orientation on RNA polymer- ase binding described in this work.

The overall shape of the promoter DNA may influence the efficiency of transcription initiation. Based on electrophoretic mobility studies, DNA containing the wild-type AM56 pro- moter is “S”-shaped. It has been proposed (49) that promoter DNA is structured by wrapping it around the RNA polymerase prior to helix untwisting and DNA strand separation. The presence of the two curves in the S-shaped promoter DNA may facilitate DNA wrapping and allow the RNA polymerase to stably contact both sides of the DNA helix.

The positioning of the runs of adenines and thymines immediately upstream of the Ah156 promoter would deflect the DNA toward the side of the helix to which the majority of the interactions between the major B. subtilis RNA polym- erase and other promoters have been mapped (47, 48). This suggests that the enhanced promoter function is dependent on the interaction between the RNA polymerase and the upstream DNA either directly (19, 22) or through a second RNA polymerase bound to the region (49). In each case, the orientation of the curve would be predicted to be critical. Several lines of evidence are inconsistent with the involve- ment of a second RNA polymerase molecule in the efficient transcription of the AM56 promoter. Although nucleotide sequence analysis identified a site with fair homology with the consensus promoter sequence located upstream of the Alu156 promoter, no significant RNA polymerase binding was observed to DNA fragments containing just this region (data not shown). When the upstream DNA of the Alu156 promoter was replaced with the analogous DNA from the Ball29 pro- moter, most of the promoter function was retained (26). The upstream DNA of the Ball29 promoter exhibited DNA cur- vature but has no secondary binding site based on nucleotide sequence analysis or RNA polymerase binding measured by the nitrocellulose filter assay. In addition, the shape of the curve in Fig. 5 does not support a model predicting the cooperative binding of more than one enzyme molecule. In other studies, chemical and enzymatic probes were used to demonstrate a secondary RNA polymerase binding site im- mediately upstream of the productive B. subtilis ueg promoter (48). Although transcription from the promoter was slightly increased in the presence of the secondary binding site, effec- tive transcription was still observed when the second site was deleted (50).

If the RNA polymerase molecule bound to the promoter makes direct contact with the upstream curved DNA, one would expect that not only the rotational orientation but also the linear placement of the curved DNA to be critical to RNA polymerase binding. The insertion of 21 bp of DNA between the -35 region and the curved DNA did not appreciably disrupt RNA polymerase binding. This suggests that the curve may loop the DNA and permit the RNA polymerase to also bind DNA upstream of the curve. Linear displacement of the curve would be less of a factor if the function of the curve is to mediate the looping of the DNA back around the RNA polymerase. If the DNA curve facilitates DNA looping up- stream of the Alu156 promoter, the DNA loop must be rela- tively small. The Alu156 hybrid promoter (26), which has just 54 bp of DNA upstream of the -35 region, was bound by RNA polymerase nearly as efficiently as the wild-type Alu156 promoter.

Acknowledgments-We thank Dm. H. D. Braymer and M. R. Belas for critical reading of the manuscript.

1. 2. 3.

4. 5.

6. 7.

8.

9. 10.

11.

12.

13.

14. 15. 16.

17. 18.

19. 20.

21. 22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

REFERENCES Diekman, S. (1986) FEBS Lett. 195,53-56 Hagerman, P. J. (1986) Nature 321,449-450 Koo, H.-S., and Crothers, D. M. (1988) Proc. Natl. Acad. Sci.

Trifonov, E. N. (1985) CRC Crit. Reu. Biochern. 19, 89-106 Ulanovsky, L., Bodner, M., Trifovov, E. N., and Choder, M.

Zahn, K., and Blather, F. R. (1987) Science 236,416-422 Ross, W., Schulman, M., and Landy, A. (1982) J. Mol. Biol. 156,

505-522 Leong, J. M., Nunes-Duby, S., Lesser, C. F., Youderian, P.,

Susskind, M. M., and Landy, A. (1985) J. Biol. Chem. 260 ,

U. S. A. 85, 1763-1767

(1986) Proc. Natl. Acad. Sci. U. S. A. 83, 862-866

4468-4477 Koepsel, R., and Khan, S. (1986) Science 233 , 1316-1318 Ryder, K., Silver, S., DeLucia, A. L., Fanning, E., and Tegtmeyer,

Snyder, M., Buchman, A. R., and Davis, R. W. (1986) Nature

Liu-Johnson, H.-N., Gartenberg, M. R., and Crothers, D. M.

Buck, M., Miller, S., Drummond, M., and Dixon, R. (1986) Nature

Reitzer, L. J., and Magasanik, B. (1986) Cell 4 5 , 785-792 Mizuno, T. (1987) Gene (Amst.) 54,57-64 Gourse, R. L., de Boer, H. A., and Homura, M. (1986) Cell 44 ,

Lamond, A. T., and Travers, A. A. (1983) Nature 305, 248-250 Bauer, B. F., Kar, E. G., Elford, R. M., and Holmes, W. M. (1988)

Bossi, L., and Smith, D. M. (1984) Cell 39,643-652 Horn, G. T., and Wells, R. D. (1981) J. Biol. Chem. 2 5 6 , 2003-

Nishi, T., and Itoh, S. (1986) Gene (Amst.) 44 , 29-36 Plaskon, R. R., and Wartell, R. M. (1987) Nucleic Acids Res. 15,

Graves, M. C., and Rabinowitz, J. C. (1986) J. Biol. Chern. 261 ,

Peschke, U., Beuck, V., Bujard, H., Gentz, R., and Le Grice, S.

Banner, C. D. B., Moran, C . P., and Losick, R. (1983) J. Mol.

McAllister, C. F., and Achberger, E. C. (1988) J. Biol. Chern.

Itakura, K., Rossi, J. J., and Wallace, R. B. (1984) Annu. Rev. Biochern. 5 3 , 323-356

Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982) Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY

Sanger, F., Nicklen, S., and Coulson, A. (1977) Proc. Natl. Acad. Sci. U. S. A. 74,5463-5467

Duvall, E., Williams, D., Lovett, P., Rudolph, C., Vasantha, N., and Guyer, M. (1983) Gene (Amst.) 2 4 , 171-177

Ostroff, G. R., and Pine, J. J. (1984) Mol. Gen. Genet. 193, 299- 305

Chang, S., and Cohen, S. N. (1979) Mol. Gen. Genet. 168 , 111- 115

P. (1986) Cell 44, 719-725

324,87-89

(1986) Cell 47,995-1005

320,374-378

197-205

Gene (Amst.) 6 3 , 123-134

2009

785-796

11409-11415

(1985) J. Mol. Biol. 186 , 547-555

Biol. 168 , 351-365

263,11743-11749

Shaw, W. (1975) Methods Enzymol. 43, 737-755 34. Bradford, M. M. (1976) Anal. Biochem. 72, 248-254 35. Achberger, E. C., and Wbiteley, H. R. (1981) J. Biol. Chem. 2 5 6 ,

7424-7432 36. Spiegelman, G . B., Hiatt, W. R., and Whiteley, H. R. (1978) J.

Biol. Chern. 253, 1756-1765 37. Jones, B. B., Chan, H., Rothstein, S., Wells, R. D., and Reznikoff,

W. S. (1977) Proc. Natl. Acad. Sci. U. S. A. 74, 4914-4918 38. Melancon, P., Burgess, R. R., and Record, M. T., Jr. (1983)

Biochemistry 22,5169-5176 39. Birnboim, H. C., and Doly, J. (1979) Nucleic Acids Res. 7, 1513-

1523 40. Marini, J. C., Effron, P. N., Goodman, T. C., Singleton, C. K.,

Wells, R. D., Wartell, R. M., and Eglund, P. T. (1984) J. Biol. Chem. 259,8974-8979

41. Wu, H.-M., and Crothers, D. M. (1984) Nature 308,509-513 42. Achberger, E. C., and Whiteley, H. R. (1980) J. Biol. Chern. 2 5 5 ,

43. von Gabain, A., Belasco, J. G., Schottel, J. L., Chang, A. C. Y., 11957-11964

10456 Positioning of Upstream DNA and Promoter Selection

and Cohen, S. N. (1983) Proc. Natl. Acad. Sci. U. S. A. 80, 653-657

44. Laundon, C. H., and Griffith, J. D. (1988) Cell 52,545-549 45. Trifonov, E. N., and Sussman, J. L. (1980) Proc. Natl. Acad. Sci.

46. Travers, A. A., Lamond, A. I., Mace, H. A. F., and Berman, M. U. S. A. 77,3816-3820

L. (1983) Cell 35, 265-273

47. Achberger, E. C., Hilton, M. D., and Whiteley, H. R. (1982)

48. LeGrice, S. F. J., and Sonenshein, A. L. (1982) J. Mol. Biol. 162,

49. Travers, A. A. (1987) CRC Crit. Rev. Biochem. 22, 181-219 50. Le Grice, S. F. J., Shih, C.-C., Whipple, F., and Sonenshein, A.

Nucleic Acids Res. 10, 2893-2910

551-564

L. (1986) Mol. Gen. Genet. 204, 229-236

Alu156- +6-

E x t -

2 3 4 5 6

.I" . 2 ',