Portal de Periódicos da CAPES

RNase H Overproduction Corrects a Defect at the Level of Transcription Elongation during rRNA Synthesis in the Absence of DNA Topoisomerase I in Escherichia coli

2000; Elsevier BV; Volume: 275; Issue: 15 Linguagem: Inglês

10.1074/jbc.275.15.11257

1083-351X

Chadi Hraiky, Marc-André Raymond, Marc Drolet,

Antibiotic Resistance in Bacteria

It has been suggested that the major function of DNA topoisomerase I in Escherichia coli is to suppress the formation of R-loops, which could inhibit growth. Although the currently available data suggest that the inhibitory effect of R-loops is exerted at the level of gene expression, this has never been demonstrated. In the present report, we show that rRNA synthesis is significantly impaired at the level of transcription elongation in a bacterial strain lacking DNA topoisomerase I. We found that this inhibition is due to transcriptional blocks. RNase H overproduction is also shown to considerably reduce the extent of such transcriptional blocks during rRNA synthesis. Moreover, one of these transcriptional blockage sites is located within a region where extensive R-loop formation was previously shown to occur on a plasmid DNA in the absence of DNA topoisomerase I. Together, these results allow us to propose that an important function of DNA topoisomerase I is to inhibit the formation of R-loops, which may otherwise translate into roadblocks for RNA polymerases. Our results also highlight the potential regulatory role of DNA supercoiling at the level of transcription elongation. It has been suggested that the major function of DNA topoisomerase I in Escherichia coli is to suppress the formation of R-loops, which could inhibit growth. Although the currently available data suggest that the inhibitory effect of R-loops is exerted at the level of gene expression, this has never been demonstrated. In the present report, we show that rRNA synthesis is significantly impaired at the level of transcription elongation in a bacterial strain lacking DNA topoisomerase I. We found that this inhibition is due to transcriptional blocks. RNase H overproduction is also shown to considerably reduce the extent of such transcriptional blocks during rRNA synthesis. Moreover, one of these transcriptional blockage sites is located within a region where extensive R-loop formation was previously shown to occur on a plasmid DNA in the absence of DNA topoisomerase I. Together, these results allow us to propose that an important function of DNA topoisomerase I is to inhibit the formation of R-loops, which may otherwise translate into roadblocks for RNA polymerases. Our results also highlight the potential regulatory role of DNA supercoiling at the level of transcription elongation. nucleotide(s) Escherichia coli DNA topoisomerase I, a member of the type IA family of topoisomerases, specifically relaxes negatively supercoiled DNA (1.Wang J.C. J. Mol. Biol. 1971; 55: 523-533Crossref PubMed Scopus (530) Google Scholar, 2.Wang J.C. Annu. Rev. Biochem. 1996; 65: 635-692Crossref PubMed Scopus (2097) Google Scholar). This specificity is explained by the fact that this enzyme binds to the junction of single-stranded and double-stranded DNA regions. DNA opening, and hence the generation of single-stranded DNA regions, is promoted by negative but not positive supercoiling. Hot spots for relaxation by DNA topoisomerase I are provided during transcription elongation in the frame of the twin-domain model (3.Liu L.F. Wang J.C. Proc. Natl. Acad. Sci. U. S. A. 1987; 84: 7024-7027Crossref PubMed Scopus (1576) Google Scholar). Indeed, very high levels of negative supercoiling can be generated behind the moving RNA polymerase during transcription elongation (4.Wu H.Y. Shyy S.H. Wang J.C. Liu L.F. Cell. 1988; 53: 433-440Abstract Full Text PDF PubMed Scopus (579) Google Scholar, 5.Tsao Y.P. Wu H.-Y. Liu L.F. Cell. 1989; 56: 111-118Abstract Full Text PDF PubMed Scopus (285) Google Scholar). An R-loop, in which the template strand is paired with the nascent RNA, leaving the nontemplate strand unpaired, also provides a hot spot for relaxation by this enzyme (6.Phoenix P. Raymond M.-A. Massé E. Drolet M. J. Biol. Chem. 1997; 272: 1473-1479Abstract Full Text Full Text PDF PubMed Scopus (52) Google Scholar). The accumulated evidence over the last few years has allowed us to conclude that a major function of DNA topoisomerase I in E. coli is to inhibit R-loop formation during transcription elongation. Indeed, the growth problem of topA (encoding DNA topoisomerase I) null mutants was shown to be partially corrected by overproducing RNase H, an enzyme that degrades the RNA moiety of an R-loop (7.Drolet M. Phoenix P. Menzel R. Massé E. Liu L.F. Crouch R.J. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 3526-3530Crossref PubMed Scopus (206) Google Scholar). A correlation was also established between the level of DNA gyrase activity, the enzyme that introduces negative supercoiling within the chromosomal DNA, and the amount of RNase H required to stimulate the growth of topA null mutants (7.Drolet M. Phoenix P. Menzel R. Massé E. Liu L.F. Crouch R.J. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 3526-3530Crossref PubMed Scopus (206) Google Scholar) and to inhibit R-loop formation during transcription (8.Massé E. Drolet M. J. Biol. Chem. 1999; 274: 16659-16664Abstract Full Text Full Text PDF PubMed Scopus (112) Google Scholar). The finding that severaltopA null mutants carry compensatory gyrmutations (in gyrA or gyrB) that reduce DNA gyrase activity and correct their growth defect (9.DiNardo S. Voelkel K.A. Sternglanz R. Reynolds A.E. Wright A. Cell. 1982; 31: 43-51Abstract Full Text PDF PubMed Scopus (295) Google Scholar, 10.Pruss G.J. Manes S.H. Drlica K. Cell. 1982; 31: 35-42Abstract Full Text PDF PubMed Scopus (243) Google Scholar) was therefore explained by the supercoiling activity of DNA gyrase, which promotes R-loop formation (7.Drolet M. Phoenix P. Menzel R. Massé E. Liu L.F. Crouch R.J. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 3526-3530Crossref PubMed Scopus (206) Google Scholar). On the contrary, DNA topoisomerase I activity inhibits R-loop formation. The results of in vitroexperiments very well support this model in which DNA topoisomerases with opposing enzymatic activities control the formation of growth-inhibitory R-loops (6.Phoenix P. Raymond M.-A. Massé E. Drolet M. J. Biol. Chem. 1997; 272: 1473-1479Abstract Full Text Full Text PDF PubMed Scopus (52) Google Scholar, 11.Drolet M. Bi X. Liu L.F. J. Biol. Chem. 1994; 269: 2068-2074Abstract Full Text PDF PubMed Google Scholar, 12.Massé E. Phoenix P. Drolet M. J. Biol. Chem. 1997; 272: 12816-12823Abstract Full Text Full Text PDF PubMed Scopus (69) Google Scholar). More recent data have suggested that local supercoiling generated during transcription, rather than global supercoiling, which reflects the average superhelical density of all supercoiling domains, is responsible for R-loop formation and hence is linked to the essential function of DNA topoisomerase I (8.Massé E. Drolet M. J. Biol. Chem. 1999; 274: 16659-16664Abstract Full Text Full Text PDF PubMed Scopus (112) Google Scholar, 13.Massé E. Drolet M. J. Biol. Chem. 1999; 274: 16654-16658Abstract Full Text Full Text PDF PubMed Scopus (61) Google Scholar). The mechanism(s) by which R-loop formation exerts its growth inhibitory effects is still unknown. The fact that topA null mutants are sensitive to changes in environmental conditions (7.Drolet M. Phoenix P. Menzel R. Massé E. Liu L.F. Crouch R.J. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 3526-3530Crossref PubMed Scopus (206) Google Scholar, 14.Dorman C.J. Lynch A.S. Ni Bhriain N. Higgins C.F. Mol. Microbiol. 1989; 3: 531-540Crossref PubMed Scopus (48) Google Scholar, 15.Qi H. Menzel R. Tse-Dinh Y.C. Mol. Microbiol. 1996; 21: 703-711Crossref PubMed Scopus (29) Google Scholar, 16.Qi H. Menzel R. Tse-Dinh Y.C. J. Mol. Biol. 1997; 267: 481-489Crossref PubMed Scopus (38) Google Scholar, 17.Massé E. Drolet M. J. Mol. Biol. 1999; 294: 321-332Crossref PubMed Scopus (45) Google Scholar) may suggest that gene expression is somehow impaired by R-loop formation. For example, the finding that RNase H overproduction allowstopA null mutants to more rapidly adapt to fresh media and to nutritional shift-ups, may suggest that R-loops inhibit the expression of genes required for such growth transitions (7.Drolet M. Phoenix P. Menzel R. Massé E. Liu L.F. Crouch R.J. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 3526-3530Crossref PubMed Scopus (206) Google Scholar, 17.Massé E. Drolet M. J. Mol. Biol. 1999; 294: 321-332Crossref PubMed Scopus (45) Google Scholar). Interestingly, we have shown previously that R-loop formation can occur during transcription of a DNA fragment carrying a portion of therrnB operon on a plasmid DNA, in the absence of DNA topoisomerase I (12.Massé E. Phoenix P. Drolet M. J. Biol. Chem. 1997; 272: 12816-12823Abstract Full Text Full Text PDF PubMed Scopus (69) Google Scholar). It has also been shown that it is during nutritional shift-up conditions that E. coli cells have the highest requirement for rRNA synthesis (encoded byrrn operons; Ref. 18.Condon C. Liveris D. Squires C. Schwartz I. Squires C.L. J. Bacteriol. 1995; 177: 4152-4156Crossref PubMed Scopus (171) Google Scholar). In this report, we present data suggesting that R-loop formation inhibits rRNA synthesis at the level of transcription elongation. Our results allow us to propose that DNA topoisomerase I can act as a transcription elongation factor that inhibits R-loop-dependent transcriptional blocks. These findings can have an important impact on our understanding of the mechanism(s) by which DNA topoisomerases and DNA supercoiling influence gene expression. E. coli strains used are listed in Table I. Details of their construction by transduction using P1vir phage are also provided in Table I. pSK760 is a pBR322 derivative carrying thernhA gene encoding RNase H (22.Kanaya S. Crouch R.J. J. Biol. Chem. 1983; 258: 1276-1281Abstract Full Text PDF PubMed Google Scholar).Table IE. coli strains used in this studyStrainGenotypeRef./constructionAQ634ilv, metB, his-29, trpA9605, pro, thyA, deoB (or C)19.Masai H. Asai T. Kubota Y. Arai K. Kogoma T. EMBO J. 1994; 13: 5338-5345Crossref PubMed Scopus (93) Google ScholarCAG18592zie-3163::Tn10kan20.Singer M. Baker T.A. Schnitzler G. Deischel S.M. Goel M. Dove W. Jaacks K.J. Grossman A.D. Erickson J.W. Gross C.A. Microbiol. Rev. 1989; 53: 1-24Crossref PubMed Google ScholarRFM445gyrB221(couR)gyrB203(Ts)7.Drolet M. Phoenix P. Menzel R. Massé E. Liu L.F. Crouch R.J. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 3526-3530Crossref PubMed Scopus (206) Google ScholarPH326RFM445zie-3163::Tn10kanRFM445XP1.CAG18592, select Kmr and TsMA249AQ634gyrB221(couR) gyrB203(Ts)zie-3163::Tn10kanAQ634XP1.PH326, select Kmr and TsRFM480topA20::Tn107.Drolet M. Phoenix P. Menzel R. Massé E. Liu L.F. Crouch R.J. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 3526-3530Crossref PubMed Scopus (206) Google ScholarMA251MA249topA20::Tn10MA249XP1.RFM480, select Tetr and cold sensitivity (27 °C)K37garB10, fhuA22, phoA4, ompF627, serU132, fadL701, relA1, pit-10, spoT1, rrnB-2, mcrB1, creC51021.Friedman D.I. Baumann M. Baron L.S. Virology. 1976; 73: 119-127Crossref PubMed Scopus (66) Google ScholarK450K37nusB521.Friedman D.I. Baumann M. Baron L.S. Virology. 1976; 73: 119-127Crossref PubMed Scopus (66) Google Scholar Open table in a new tab Unless otherwise indicated, the strains were grown in Vogel-Bonner minimal medium (7.Drolet M. Phoenix P. Menzel R. Massé E. Liu L.F. Crouch R.J. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 3526-3530Crossref PubMed Scopus (206) Google Scholar) supplemented with the required amino acids (50 μg/ml), casamino acids (0.1%), glucose (0.2%), thiamine (5 μg/ml), and, when indicated, thymine at 10 μg/ml. When needed, ampicillin was added to 50 μg/ml, tetracycline to 10 μg/ml, and kanamycin to 27.5 μg/ml. Overnight cultures were diluted 1/50 in fresh medium supplemented with cold thymidine at 1 μg/ml and then grown to an A 600 of 0.4. Cells were transferred to 27 °C and incubated for the times indicated in the legends to figs. 2, 6, 7, 8, and 9. Pulse labeling was initiated by adding 24 μCi of [3H]uridine (stock solution at 57 Ci/mmol) to 1.8 ml of cell culture. Total RNA was extracted from 600 μl of cell culture after 3, 5, and 8 min of pulse labeling by a modified version of the hot-phenol procedure of Aiba et al. (23.Aiba H. Adhya S. de Crombrugghe B. J. Biol. Chem. 1981; 256: 11905-11910Abstract Full Text PDF PubMed Google Scholar). Instead of being recovered by centrifugation, the cells were directly treated with the lysis solution containing the various ingredients at the required amount to yield the final concentration used in the original protocol. The purified RNA was resuspended in 20 μl of diethyl pyrocarbonate-treated deionized water and stored at −70 °C. One μg of the RNA samples was analyzed by electrophoresis in 1.2% denaturing agarose gel containing formaldehyde as described (24.Sambrook J. Fritsch E.F. Maniatis T. Molecular Cloning: A Laboratory Manual . Cold Spring Harbor Laboratory, Cold Spring Harbor, NY1989Google Scholar). After electrophoresis, the gel was photographed under UV light. In order to reveal the 3H-labeled rRNA, the gels were treated for fluorography with EN3HANCE (NEN Life Science Products), dried, and exposed to x-ray films (RX, Fuji). One μg of the RNA samples was used for primer extension by using avian myeloblastosis virus reverse transcriptase (Roche Diagnostic) at 42 °C according to published procedures (24.Sambrook J. Fritsch E.F. Maniatis T. Molecular Cloning: A Laboratory Manual . Cold Spring Harbor Laboratory, Cold Spring Harbor, NY1989Google Scholar). Primer d(GGGTGTGCATAATACGCC) was used in all the experiments; it allows the detection of RNA initiated at the P1 promoter of all E. coli rrn operons. The products of the primer extension experiments were precipitated and loaded on 8% denaturing polyacrylamide gels. RNA samples extracted as described above were loaded on 1.2% denaturing agarose gels containing formaldehyde (24.Sambrook J. Fritsch E.F. Maniatis T. Molecular Cloning: A Laboratory Manual . Cold Spring Harbor Laboratory, Cold Spring Harbor, NY1989Google Scholar). After electrophoresis, the gels were photographed under UV light. RNA was transferred to a nylon membrane, and hybridization took place according to published procedures (24.Sambrook J. Fritsch E.F. Maniatis T. Molecular Cloning: A Laboratory Manual . Cold Spring Harbor Laboratory, Cold Spring Harbor, NY1989Google Scholar). The probe (138 base pairs), covering a portion of the leader rRNA, was obtained by polymerase chain reaction using pNO1302 DNA (25.Jinks-Robertson S. Gourse R.L. Nomura M. Cell. 1983; 33: 865-876Abstract Full Text PDF PubMed Scopus (143) Google Scholar) and two oligonucleotides with the sequences d(GGAACAACGGCAAACACG) and d(GCCGCTTCGCTTTTTCTC). Following hybridization, the nylon membranes were washed according to conventional procedures (24.Sambrook J. Fritsch E.F. Maniatis T. Molecular Cloning: A Laboratory Manual . Cold Spring Harbor Laboratory, Cold Spring Harbor, NY1989Google Scholar) and exposed to x-ray films (RX, Fuji). To evaluate the impact of the loss of DNA topoisomerase I activity on rRNA synthesis, two isogenic strains carrying a Ts gyrB allele were used. One of these strains, MA251, also carries thetopA20::Tn10 allele and is therefore totally deprived of topoisomerase I activity. The combination of these two alleles was used previously to reveal the link between R-loop formation and growth inhibition in the absence of DNA topoisomerase I (7.Drolet M. Phoenix P. Menzel R. Massé E. Liu L.F. Crouch R.J. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 3526-3530Crossref PubMed Scopus (206) Google Scholar). Indeed, such a topA null mutant is able to grow at 37 °C because DNA gyrase activity is low enough to compensate for the absence of DNA topoisomerase I, therefore reducing R-loop formation to a tolerable level for the cell. As the temperature decreases, DNA gyrase becomes more active in promoting R-loop formation (8.Massé E. Drolet M. J. Biol. Chem. 1999; 274: 16659-16664Abstract Full Text Full Text PDF PubMed Scopus (112) Google Scholar). RNase H overproduction was shown to significantly improve the growth of thistopA null mutant below 37 °C, down to 27 °C (7.Drolet M. Phoenix P. Menzel R. Massé E. Liu L.F. Crouch R.J. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 3526-3530Crossref PubMed Scopus (206) Google Scholar). To measure the synthesis rate of mature 16 S rRNA and 23 S rRNA, total RNA was pulse-labeled for short periods of time with [3H]uridine. Such experiments could reveal defects at the level of transcription and/or at the level of rRNA maturation. The ratio of the promoter distal 23 S rRNA to the promoter proximal 16 S rRNA (see Fig. 1 for a schematic drawing of an rrn operon) synthesized in these experiments could also reveal problems during transcription elongation and/or differences in the rate of 16 S and 23 S rRNA maturation. The pulse labeling of total RNA of cells growing at 37 °C revealed no significant differences in rRNA synthesis rates, nor in the ratio of the distal (23 S) to proximal (16 S) rRNA, between MA249 (gyrB(Ts)) and MA251 (gyrB(Ts)topA20::Tn10) (data not shown). The next series of pulse-labeling experiments was performed with cells exposed to 27 °C, a nonpermissive temperature for the growth of MA251. In order to show that our approach can indeed be used to reveal problems at the level of transcription elongation, we included a pair of isogenic strains, K37 and K450. The latter carries thenusB5 mutation and was shown previously to be defective in antitermination of rRNA transcription (26.Sharrock R.A. Gourse R.L. Nomura M. Proc. Natl. Acad. Sci. U. S. A. 1985; 82: 5275-5279Crossref PubMed Scopus (59) Google Scholar). One consequence of such a defect is an increase in both rRNA transcription initiation and in the rate of 16 S rRNA synthesis (26.Sharrock R.A. Gourse R.L. Nomura M. Proc. Natl. Acad. Sci. U. S. A. 1985; 82: 5275-5279Crossref PubMed Scopus (59) Google Scholar). Pulse labelings were performed for 3, 5, and 8 min. In our hands, shorter pulse-labeling times were shown to be insufficient to reveal a significant amount of labeled mature rRNAs (data not shown). This indicates that any difference observed in the amounts of mature rRNA synthesized following a 3-min pulse labeling cannot be attributed to mature rRNA degradation but must rather be attributed to alterations in rRNA synthesis rates (transcription and/or maturation). K37 and K450 cells were grown at 37 °C to anA 600 of 0.4, at which time they were exposed to 27 °C for 20 min. Pulse labelings with [3H]uridine for 3, 5, and 8 min were then performed. As shown previously (26.Sharrock R.A. Gourse R.L. Nomura M. Proc. Natl. Acad. Sci. U. S. A. 1985; 82: 5275-5279Crossref PubMed Scopus (59) Google Scholar), our results indicate that the 16 S rRNA synthesis rate is higher in thenusB5 mutant as compared with the rate in the isogenicnusB + strain (Fig.2 B, pulse-labeled rRNA and 2C, ratio of 23 S rRNA to 16 S rRNA). Our experimental approach is therefore appropriate to reveal defects in rRNA synthesis at the level of transcription elongation. Results presented in Fig. 2 also show that the rRNA synthesis rate is significantly reduced when thetopA null mutant is exposed to 27 °C (compare MA249 and MA251). Densitometry analysis reveals that the topA null mutant (MA251) synthesized at least three times less 23 S rRNA after a 5-min pulse labeling compared with the isogenic MA249 strain (data not shown). It can also be seen that the ratio of 23 S to 16 S rRNA synthesis rates is significantly reduced in the topA null mutant (Fig. 2 C, compare MA249 and MA251). Fig. 2 also shows that RNase H overproduction in the topA null mutant (MA251/pSK760) restores both the rRNA synthesis rate (Fig.2 B, compare MA251 and MA251/pSK760) and the 23 S to 16 S ratio (Fig. 2 C, compare MA251 and MA251/pSK760). These results suggest that R-loop formation during rRNA synthesis causes a defect at the level of transcription initiation or elongation and/or at the level of rRNA maturation. Primer extension analysis with an oligonucleotide (Fig. 1) annealing to the usually unstable leader rRNA (27.Sarmientos P. Sylvester J.E. Contente S. Cashel M. Cell. 1983; 32: 1337-1346Abstract Full Text PDF PubMed Scopus (168) Google Scholar, 28.Gafny R. Cohen S. Nachaliel N. Glaser G. J. Mol. Biol. 1994; 243: 152-156Crossref PubMed Scopus (31) Google Scholar, 29.Aviv M. Giladi H. Oppenheim A.B. Glaser G. FEMS Microbiol. Lett. 1996; 140: 71-76Crossref PubMed Google Scholar) reveals that the drop in the rRNA synthesis rate in thetopA null mutant cannot be attributed to a reduction in transcription initiation. In fact, results presented in Fig.3 may suggest that transcription initiation is increased in the topA null mutant not overproducing RNase H (compare MA249, MA251, and MA251/pSK760). As shown before (26.Sharrock R.A. Gourse R.L. Nomura M. Proc. Natl. Acad. Sci. U. S. A. 1985; 82: 5275-5279Crossref PubMed Scopus (59) Google Scholar), it can also be seen that rRNA transcription initiation is increased in the nusB5 mutant (Fig. 3, compare K37 and K450). In agreement with the negative feedback model for the regulation of rRNA synthesis, rRNA transcription is initiated more frequently to compensate for a defect at the level of transcription elongation in the nusB5 mutant (26.Sharrock R.A. Gourse R.L. Nomura M. Proc. Natl. Acad. Sci. U. S. A. 1985; 82: 5275-5279Crossref PubMed Scopus (59) Google Scholar, 30.Gourse R.L. Gaal T. Bartlett M.S. Appleman J.A. Ross W. Annu. Rev. Microbiol. 1996; 50: 645-677Crossref PubMed Scopus (209) Google Scholar). A similar hypothesis could also be made to explain the results obtained with thetopA null mutant. This question is addressed in the next section.Figure 2rRNA synthesis rate in a topAnull mutant at a nonpermissive temperature. Cells were grown at 37 °C to an A 600 of 0.4 as indicated under “Experimental Procedures,” transferred to 27 °C, and incubated for 10 (MA249), 20 (K37 and K450), or 40 min (MA251 and MA251/pSK760). RNA was next pulse-labeled at the same temperature for the indicated time and extracted as described under “Experimental Procedures.” One μg of the RNA samples was analyzed by electrophoresis in 1.2% denaturing agarose gel containing formaldehyde as described under “Experimental Procedures.” A, ethidium bromide-stained gel photographed under UV light. B, autoradiography of the dried gel after its treatment for fluorography. C, diagram showing the molar ratio of 23 S to 16 S rRNA. The values were obtained following quantification by densitometry and by taking into account the number of U residues within 16 S and 23 S rRNA. The 23 S to 16 S rRNA ratio obtained for MA249 reflects very well the situation in this genetic background because an identical ratio was obtained for the wild-type isogenic strain AQ634 (data not shown). The results shown here are representative of four independent experiments.View Large Image Figure ViewerDownload (PPT)Figure 3Primer extension analysis to detect the RNAs initiated at the rrn P1 promoter. One μg of the RNA samples from time 3 min of the experiment shown in Fig. 2 was used for primer extension analysis with a primer hybridizing to the leader rRNA, as described under “Experimental Procedures.” The results shown here are representative of four independent experiments.View Large Image Figure ViewerDownload (PPT) Our pulse-labeling experiments only measured the synthesis rate of mature rRNAs. In E. coli cells, several steps, all of which are completed very rapidly, are required to process the large rRNA precursors into mature rRNA products (31.Srivastava A.K. Schlessinger D. Annu. Rev. Microbiol. 1990; 44: 105-129Crossref PubMed Scopus (184) Google Scholar). In normal cells, only 1–2% of rRNA is present as long precursors. A reduction in the mature rRNA synthesis rate could possibly be explained by a defect in one or more of the several steps required for full rRNA maturation. However, if this were happening, large rRNA precursors should be easily detected, which is not the case here for the topA null mutant (Fig. 2and data not shown). The first step in the maturation process is the removal of the leader rRNA (180 nucleotides long) that occurs once the 16 S rRNA sequence has been fully transcribed. This step is triggered by the formation of an appropriate secondary structure involving both a 5′ end and a 3′ end portion of the 16 S rRNA precursor (Fig.4). Cleavage by RNase III follows and both the leader rRNA and the 16 S rRNA precursor are released. The leader rRNA is rapidly degraded, and the large 16 S rRNA precursor is quickly maturated. Therefore, in order to search for prematurely terminated rRNAs or for transcriptional blocks (which reflect blocked RNA polymerases), it was necessary to use a DNA probe covering no more than the leader rRNA region. In this manner, either not entirely transcribed 16 S rRNA or the cleaved leader rRNA could be detected by Northern blot analysis (Fig. 4). Aliquots of total RNA extracted for the experiments shown in Fig. 2 and3 were probed with a DNA fragment covering a portion of the leader rRNA (see under “Experimental Procedures”) in a Northern blot experiment. Results shown in Fig. 5reveal the presence of at least four different hybridization signals. One of them, the shortest RNA detected, with roughly 180 nucleotides, probably corresponds to the leader rRNA region. The amount of this short RNA species detected in each strain matches very well with the intensity of the primer extension signal revealed by using an oligonucleotide complementary to a portion of the leader rRNA region (Fig. 3). The longest RNA detected, roughly 2000 nt1 in length, may represent unprocessed 16 S rRNA that was prematurely terminated. This rRNA is easily detected in the nusB5 mutant that was previously shown to be defective in transcription antitermination during rRNA synthesis (26.Sharrock R.A. Gourse R.L. Nomura M. Proc. Natl. Acad. Sci. U. S. A. 1985; 82: 5275-5279Crossref PubMed Scopus (59) Google Scholar). Two significant hybridization signals, corresponding roughly to 550- and 1200-nt RNAs in size, were detected in RNA samples extracted from topA null mutants, irrespective of the level of RNase H activity (MA251 and MA251/pSK760). These RNA species were also detected, albeit in much smaller amounts, when the topAnull mutants were grown at 37 °C (data not shown). Interestingly, the 3′ end of the shorter one (550 nt) is situated within a region where extensive R-loop formation was previously shown to occur upon transcription on a plasmid DNA in the absence of DNA topoisomerase I (12.Massé E. Phoenix P. Drolet M. J. Biol. Chem. 1997; 272: 12816-12823Abstract Full Text Full Text PDF PubMed Scopus (69) Google Scholar). These two RNAs most likely reflect prematurely terminated RNAs or transcriptional blocks (Fig. 4). Our primer extension analysis with RNA extracted from topAnull mutants not overproducing RNase H shows that the leader rRNA signal is stronger when the RNA is extracted from cells exposed for longer periods of time to 27 °C (data not shown). This can be taken as an indication that one or several stable RNA(s) are accumulating with time. This could reflect transcriptional blocks with the 3′ RNA ends within blocked RNA polymerases being protected from degradation, rather than prematurely terminated RNAs, with freed 3′ ends that are rapidly degraded (Fig. 4). Moreover, when a Northern blot experiment is performed with RNA samples extracted from cells exposed to 27 °C for various lengths of time, it can be seen that the distribution of the hybridization signals is modified (Fig.6). Indeed, it is apparent that upon longer exposition of the cells to 27 °C, the hybridization signals are shifted toward the shorter RNAs, and therefore toward the 5′ proximal region. Similar results were also obtained with a bacterial strain in which the topA gene is deleted (RFM475, Ref. 7.Drolet M. Phoenix P. Menzel R. Massé E. Liu L.F. Crouch R.J. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 3526-3530Crossref PubMed Scopus (206) Google Scholar; data not shown). It is also obvious that this shift is more severe when the topA null mutant do not overproduce RNase H. This may indicate that stable transcriptional blocks are accumulating with time when the level of RNase H activity is too low. The results of experiments described below suggest that this interpretation is most likely correct. To discriminate between transcriptional blocks and prematurely terminated RNAs, we measured the stability of the RNA species carrying the leader rRNA region. RNA stability was measured by adding the transcription initiation inhibitor rifampicin to cells previously exposed to 27 °C for 40 min and thereafter by periodically extracting RNA samples. Primer extension analysis (Fig.7) evaluates the stability of the whole population of RNAs carrying the leader rRNA region, whereas a Northern blot analysis (Fig. 8) measures the stability of every one of these RNAs individually. First of all, the primer extension analysis reveals that the RNA population carrying the leader rRNA in the gyrB(Ts), topA +strain (MA249) is very unstable, its half-life being less than 1 min, as previously shown for the cleaved leader rRNA (Fig. 7 and data not shown; Refs. 27.Sarmientos P. Sylvester J.E. Contente S. Cashel M. Cell. 1983; 32: 1337-1346Abstract Full Text PDF PubMed Scopus (168) Google Scholar, 28.Gafny R. Cohen S. Nachaliel N. Glaser G. J. Mol. Biol. 1994; 243: 152-156Crossref PubMed Scopus (31) Google Scholar, 29.Aviv M. Giladi H. Oppenheim A.B. Glaser G. FEMS Microbiol. Lett. 1996; 140: 71-76Crossref PubMed Google Scholar). For this strain, the primer extension analysis reflects the cleaved leader rRNA, because it is the only rRNA species detected by Northern blot analysis (Fig. 5). However, for the isogenictopA null mutant MA251, primer extension analysis shows that the RNA population carrying the leader rRNA region is very stable (Fig.7). In fact, we have found that the primer extension signal is still very strong after 20 min of rifampicin treatment (data not shown). The Northern blot analysis shown in Fig. 8 A reveals that the stable products represent three different RNA species. One has its 3′ end within the leader rRNA region (the strongest signal), and the other two have their 3′ ends situated about 550 or 1200 nt from the P1 promoter initiation site. Considering the very high stability of these RNAs, mostly the one with the 3′ end within the leader rRNA region, we have to conclude that they reflect transcriptional blocks rather than prematurely terminated RNAs. Stable RNA-DNA hybrids are probably involved in transcriptional blocks, because overproducing RNase H renders these RNAs less stable (Figs. 7 and 8, MA249/pSK760). Quantitative evaluations of the half-life of each RNA species (Fig.8 B) reveal that the RNA with the 3′ end within the leader rRNA region is at least six times more stable in the topAnull mutant when RNase H is not overproduced. For the two longer RNAs (∼1200 and 550 nt), the half-life roughly double when RNase H is not overproduced. However, these evaluations may be complicated by the fact that RNA polymerase molecules may progress from 5′ proximal block sites to more distal ones with rifampicin exposition time. The fact that the 5′ proximal block site is more stable than the more distal ones may explain the absence of smears, reflecting a piling up of RNA polymerase molecules, instead of sharp bands in the Northern blots. The results presented in Fig. 9 further illustrate this fact and the high stability of the transcriptional blocks intopA null mutants when RNase H is not overproduced. Indeed, under such conditions, strong hybridization signals were still detected after 30 min of exposition to rifampicin.Figure 8Northern blot analysis to measure the stability of each of the different RNA species initiated at therrn P1 promoter. Cells were grown at 37 °C to an A 600 of 0.4 as indicated under “Experimental Procedures” and transferred to 27 °C. After 40 min of incubation at this temperature, rifampicin at 250 μg/ml (final concentration) was added to the cells, and the RNA was extracted at the indicated time (for time 0, the RNA was extracted immediately before the addition of rifampicin). Three μg of the RNA samples were used for Northern blot analysis with a probe hybridizing to the leader rRNA, as described under “Experimental Procedures.” The top panel in A represents the ethidium bromide-stained gel photographed under UV light, and the bottom panel is an autoradiography of the Northern blot. In B, the intensity of the hybridization signals for the different RNA species was measured by densitometry to evaluate the half-life of each of these RNA species. The results shown here are representative of three independent experiments. The black line is MA251, and the gray line is MA251/pSK760.View Large Image Figure ViewerDownload (PPT)Figure 9Northern blot analysis to reveal the very high stability of the different RNA species initiated at therrn P1 promoter in topA null mutants. Cells were grown at 37 °C to anA 600 of 0.4 as indicated under “Experimental Procedures” and transferred to 27 °C. After 40 min of incubation at this temperature, rifampicin at 250 μg/ml (final concentration) was added to the cells, and the RNA was extracted at the indicated time (for time 0, the RNA was extracted immediately before the addition of rifampicin). Three μg of the RNA samples were used for Northern blot analysis with a probe hybridizing to the leader rRNA, as described under “Experimental Procedures.” The top panelrepresents the ethidium bromide-stained gel photographed under UV light, and the bottom panel is an autoradiography of the Northern blot.View Large Image Figure ViewerDownload (PPT) Although it has been known for a long time that topAnull mutations are detrimental to cell growth and that DNA supercoiling is somehow involved in this inhibition (9.DiNardo S. Voelkel K.A. Sternglanz R. Reynolds A.E. Wright A. Cell. 1982; 31: 43-51Abstract Full Text PDF PubMed Scopus (295) Google Scholar, 10.Pruss G.J. Manes S.H. Drlica K. Cell. 1982; 31: 35-42Abstract Full Text PDF PubMed Scopus (243) Google Scholar, 32.Drlica K. Mol. Microbiol. 1992; 6: 425-433Crossref PubMed Scopus (292) Google Scholar), no precise mechanism(s) has been described to explain such a negative effect. One of the first clues that R-loop formation could somehow be involved in this harmful effect came from the results of in vitrostudies in which extensive RNA-DNA hybrids were shown to occur during transcription in the presence of DNA gyrase and in the absence of DNA topoisomerase I (11.Drolet M. Bi X. Liu L.F. J. Biol. Chem. 1994; 269: 2068-2074Abstract Full Text PDF PubMed Google Scholar). A key finding was the demonstration that RNase H overproduction can partially complement the growth defect oftopA null mutants (7.Drolet M. Phoenix P. Menzel R. Massé E. Liu L.F. Crouch R.J. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 3526-3530Crossref PubMed Scopus (206) Google Scholar). It was proposed that R-loops could act by triggering inappropriate replication or recombination, such as is suggested to occur in rnhA − mutant (33.Kogoma T. Microbiol. Mol. Biol. Rev. 1997; 61: 212-238Crossref PubMed Scopus (446) Google Scholar), or by inhibiting transcription elongation. In the present work, we have shown that transcriptional blocks in the absence of DNA topoisomerase I can be removed by overproducing RNase H. Interestingly, we had previously found that RNase H stimulates the RNA synthesis rate in vitro when a negatively supercoiled template was transcribed (6.Phoenix P. Raymond M.-A. Massé E. Drolet M. J. Biol. Chem. 1997; 272: 1473-1479Abstract Full Text Full Text PDF PubMed Scopus (52) Google Scholar,12.Massé E. Phoenix P. Drolet M. J. Biol. Chem. 1997; 272: 12816-12823Abstract Full Text Full Text PDF PubMed Scopus (69) Google Scholar). 2E. Massé and M. Drolet, unpublished results. Moreover, one of the transcriptional blocks identified in the present study was found within a region where extensive R-loop formation was observed during in vitro and in vivo transcription on plasmid DNAs in the absence of DNA topoisomerase I (6.Phoenix P. Raymond M.-A. Massé E. Drolet M. J. Biol. Chem. 1997; 272: 1473-1479Abstract Full Text Full Text PDF PubMed Scopus (52) Google Scholar, 12.Massé E. Phoenix P. Drolet M. J. Biol. Chem. 1997; 272: 12816-12823Abstract Full Text Full Text PDF PubMed Scopus (69) Google Scholar). Together, these observations allow us to propose that the R-loops generated due to the absence of DNA topoisomerase I constitute roadblocks for RNA polymerases. In our model, the nascent RNA invades the template DNA strand behind a moving RNA polymerase, to form an R-loop (Fig.10, 1). This R-loop constitutes a roadblock for the next transcribing RNA polymerases (Fig.10, 2). For the moment, we can only speculate on what the real effects of such transcriptional blocks are on the growth oftopA null mutants, although RNase H overproduction strongly stimulates their growth at 27 °C (7.Drolet M. Phoenix P. Menzel R. Massé E. Liu L.F. Crouch R.J. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 3526-3530Crossref PubMed Scopus (206) Google Scholar). Moreover, because our experimental approaches did not allow to discriminate between the seven different rrn operons of E. coli, we cannot draw a conclusion about the involvement of each of these operons separately, in transcriptional blockage. The fact that a topA null mutant synthesizes at least three times less 23 S rRNA than an isogenictopA + strain during a 5-min pulse-labeling experiment leads us to believe that the observed transcriptional blocks contribute to the growth rate reduction in the absence of DNA topoisomerase I. This is also supported by our finding that the growth of a topA null mutant overproducing RNase H at 27 °C can be further improved by increasing the number of rrn operons per cell. 3M. Drolet and C. Hraiky, unpublished results. This is accomplished by introducing a multicopy plasmid carrying a fully activerrn operon (rrnB) within the topA null mutant. The transcriptional blocks identified within rrn operons seem to be very stable. This is suggested by the results of both primer extension and Northern blot analysis performed with RNA extracted from cells exposed to rifampicin. Indeed, a strong signal was still detected by Northern blot analysis with a DNA fragment hybridizing to the leader rRNA region after an exposition of 30 min to rifampicin. This high stability is also revealed by the fact that the transcriptional block sites were shifted toward the 5′ proximal region upon prolonged expositions of the topA null cells to 27 °C. This is expected because the accumulation of stable transcriptional blocks in the 5′ proximal region precludes downstream transcriptional activity, and consequently, transcriptional blockage in this downstream region. Moreover, our primer extension experiments show that the stability of the RNA species carrying the leader rRNA increases with the time of exposition of the topA null mutant to 27 °C (data not shown). This also indicates that stable transcriptional blocks are slowly accumulating with time. What could explain such a high stability? If an R-loop alone is involved in transcriptional blockage, such a high stability may seem quite surprising considering the presence of RNase H. It is also possible that the initiated R-loop, immediately behind the moving RNA polymerase, provides a single-stranded DNA region, to which RecA or single-stranded binding protein can bind. Thereafter, this binding may progress with the moving RNA polymerase. Therefore, if RNase H does not act rapidly enough, a very long and stably unwound DNA region may form. Such a structure could possibly block the progression of the next transcribing RNA polymerase. The formation of other RNase H-sensitive stable structures may also be considered. For example, transcription of a long polypurine sequence was shown to favor the formation of a very stable structure that involves the participation of an RNA-DNA hybrid and that inhibits transcription elongation (34.Grabczyk E. Fishman M.C. J. Biol. Chem. 1995; 270: 1791-1797Abstract Full Text Full Text PDF PubMed Scopus (64) Google Scholar). It will also be interesting to test whether the RNase H-sensitive stable structure can inhibit replication fork movement and, as a consequence, possibly trigger the formation of double-stranded breaks (35.Michel B. Ehrlich S.D. Uzest M. EMBO J. 1997; 16: 430-438Crossref PubMed Scopus (383) Google Scholar). In fact, we had observed previously that the duplication rate of rrn operons can be very high in some topA null mutants.2 In light of the results presented in this report and the pleiotropic effects of topA mutations (32.Drlica K. Mol. Microbiol. 1992; 6: 425-433Crossref PubMed Scopus (292) Google Scholar, 36.Drlica K. Microbiol. Rev. 1984; 48: 273-289Crossref PubMed Google Scholar, 37.Dorman C.J. Ni Bhriain N. Trends Microbiol. 1993; 1: 92-99Abstract Full Text PDF PubMed Scopus (32) Google Scholar), we must consider the possibility that transcription elongation might be frequently involved, more than previously suspected, in the control of gene expression. To our knowledge, there has been only one report that directly addressed the effect of DNA supercoiling on transcription elongation in vitro (38.Krohn M. Pardon B. Wagner R. Mol. Microbiol. 1992; 6: 581-589Crossref PubMed Scopus (24) Google Scholar). In that work, in vitro transcriptional pausing times within the leader region of the rrnB operon were shown to increase with the level of negative supercoiling of the DNA template. Moreover, new pause sites were revealed at higher negative superhelical densities of the DNA template. Whether a link can be established between these observations and the results reported in the present study remains to be demonstrated. In conclusion, we believe that care must be taken before attributing the effects of DNA supercoiling and DNA topoisomerases on gene expression to transcription initiation. More detailed studies may reveal effects at the level of transcription elongation rather than initiation. We thank Dr. D. I. Friedman for the gift of bacterial strains and Sonia Broccoli for careful reading of the manuscript.