Proc. Nadl. Acad. Sci. USA Vol. 82, pp. 4663-4667, July 1985 Biochemistry activates the paused complex and restores transcription of the trp leader region (tanscripfton-tnlation coupling/transcrIptIon pausing/attenuation/pause complex half-) ROBERT LANDICK, JANNETrE CAREY*, AND CHARLES YANOFSKY Department of Biological Sciences, Stanford University, Stanford, CA 94305 Contributed by Charles Yanofsky, April 5, 1985

ABSTRACT It has been proposed that RNA polymerase have been identified in the trp (6-11), ilv (12), and thr (13) pausing in the leader region of the (tIp) operon of operon leader regions, at the position corresponding to the 3' is responsible for the synchronization of end of the first RNA hairpin. If the paused RNA polymerase transcription and translation essential to attenuation control. remained at this site until movement caused re- In this report we use an in vitro coupled transcription/ sumption of transcription, synchronization of transcription translation system to study the effect of tip leader peptide and translation would be accomplished. synthesis on RNA polymerase pausing in the tip leader region. Previous studies with the demonstrated that (l) Wild-type and translation-defective tip leader templates of E. the predicted stability of the pause RNA hairpin is directly cofi and Sernitia marcescens were employed, and pause RNA correlated with the strength of the pause signal (10, 11); (it) synthesis and paused complex release (activation) were quan- NusA protein enhances pausing by RNA polymerase (9, 10, tified relative to synthesis of the terminated leader transcript. 14); (iih) low concentrations of GTP but not the other NTPs It was observed that pausing in the tip leader region was (guanylic acid is the next nucleotide to be added) also prolonged when translation of the leader transcript was re- enhance pausing (6, 10, 11); (iv) mutant RNA polymerases duced by mutations in the leader region or by addition of the that increase or decrease transcription termination at the trp translation inhibitor kasugamycin or chloramphenicol. Exper- have analogous effects on pausing (4, 8); and (v) iments with S-30 extracts from a mutant strain that is ineffi- RNA polymerase pausing at the trp leader pause site can be cient in translating the tryptophan codons in the leader detected in a coupled transcription/translation system (11). transcript indicated that ribosome movement to these codons In this report we use the coupled system to show that the also releases the paused transcription complex. These findings ribosome engaged in synthesizing the leader peptide releases indicate that the paused tip leader transcription complex the paused complex.! resumes transcription when released by ribosome movement over the leader peptide coding region. This release would MATERIALS AND METHODS facilitate the coupling of transcription and translation essential to attenuation control. Nucleoside triphosphates, rifampicin, kasugamycin, chlor- amphenicol, and reagents required for cell-free transcription/ Transcription attenuation regulates expression of many ami- translation reactions were purchased from Sigma. [a- no acid biosynthetic of enteric (for recent 32PJGTP and [35S]methionine were from Amersham. RNA reviews, see refs. 1-4). According to the current model of polymerase and tryptophanyl-tRNA synthetase were gener- attenuation, formation ofalternative secondary structures in ous gifts of R. Fisher and K. Muench, respectively. Plasmid the transcript of the leader region controls transcription DNAs used as templates for S-30 reactions were prepared by termination at a site immediately preceding the structural standard procedures (15, 16). Restriction fragments used for genes of the operon. It is believed that translation of a short in vitro transcription reactions were prepared by Sau3A peptide coding region in the leader transcript governs for- (Escherichia coli templates) or Hpa II (Serratia marcescens mation of these alternative transcript secondary structures. templates) digestion of appropriate plasmids (17, 18, 19) and In the trp operons ofall enteric bacteria examined the leader isolated from polyacrylamide gels. S-30 cell-free extracts peptide coding region contains tandem tryptophan codons were prepared from RNaseI- strains as described (20, 21, 22). (5). Ribosome stalling at either ofthese tryptophan codons is In Vitro Transcription (Noncoupled). Paused complex half- believed to promote formation of a transcript secondary were determined in synchronized single-round in vitro structure, the antiterminator, that prevents transcription transcription experiments conducted as described (6, 9, 10). termination in the leader region. However, when a ribosome Paused complex half-lives were calculated by linear regres- translates the entire leader peptide coding region without sion analysis of the logarithm of pause RNA concentrations stalling it promotes formation of a different transcript sec- vs. time. ondary structure, the . This secondary structure Coupled Transcription/Translation Reactions. Cell-free causes transcription termination at the attenuator in the transcription/translation experiments were performed essen- leader region. The attenuation model demands that transla- tially as described (20). To measure RNA synthesis unlabeled tion of the leader peptide coding region of the transcript be GTP was replaced by [a-32P]GTP. To measure protein closely coupled to transcription of the leader region. This synthesis unlabeled methionine was replaced by [35S]methi- coupling could be achieved if the polymerase molecule onine. Plasmid DNAs were added to a final concentration of transcribing the leader region paused until it was approached 40 nM. The translation inhibitors chloramphenicol and by the ribosome engaged in translating the leader peptide coding region of the transcript. Transcription pause sites *Present address: Department ofBiochemistry, Stanford University Medical School, Stanford, CA 94305. tIn this paper we use "release of the paused complex" or "paused The publication costs ofthis article were defrayed in part by page charge complex release" to mean that RNA polymerase has resumed payment. This article must therefore be hereby marked "advertisement" transcription elongation after having paused at the trp leader pause in accordance with 18 U.S.C. ยง1734 solely to indicate this fact. site.

4663 Downloaded by guest on October 2, 2021 4664 Biochemistry: Landick et aL Proc. Natl. Acad Sci. USA 82 (1985)

kasugamycin were added to the cell-free extract prior to 1:2 A A mixing with the other reaction components. This order of A A STOP G=C mixing was necessary to achieve even 50%o inhibition of U -G C=G protein synthesis by kasugamycin. Reactions were initiated AAGUUCACG SER C-G by adding the cell-free extract (20% of final volume) and A U-A A THR C=G shifting the mixture to 370C. After 15 min of incubation to A A-U A C=G allow transcription and translation to reach steady-state A ARG G U G c A activity, [a-32P]GTP was added. In time course experiments, U G G U after a 2-min labeling period rifampicin was added to block U TRP G=C A U-A further transcription initiation. Samples were removed and U TRP G=C added to 0.6 vol of phenol at the times indicated. To measure C METLYSALAILEPHEVALLEULYSGLYG=C GACAAUAAAGCAAUUUUCGUACUGAAAGGUU-AU * () steady-state pause RNA levels, the reactions were stopped A with 0.6 vol of phenol after the 2-min labeling period. After T RPL29 centrifugation, the aqueous layers of samples were added to an equal volume of transcription stop mix (23) and electro- B phoresed on 10% polyacrylamide gels containing 7 M urea. A310--lA'.'I A C The relative pause RNA level was determined by counting A G the Cerenkov radiation of gel slices and calculating the molar C=G percent of pause RNA in the total of pause plus leader RNAs C=f=c- -A221- = A-b 1:2 (relative pause RNA level molar percent of pause RNA in AAGUUC=GUCGGAU LEU the total ofpause plus leader RNAs). The relative pause RNA CUG LEU U C level is an arbitrary steady-state measure of the half-life for C=G ARG C=G release of paused transcription complexes and is dependent SER C=G on the length of the labeling period (2 min for all measure- U C-UC ALA IHR C-U ments). Because of the complex kinetics of transcription in A-U VAL the coupled system, we used the relative pause RNA level AUG calculation to simplify analysis of coupled transcription/ ARG A STOP measure A translation pausing data. One complication of this is TRP GUC that transcription read-through at the attenuator influences METASNTHRTYRILESERLEUHISGLYTRP C-G the amount of leader RNA. By using a template containing a GUCUGCAAAUGAACACAUACAUUUCUCUUCACGGUUGGUGG=C (G) strong transcription terminator downstream from the trp FIG. 1. Sequence and secondary structure of trp leader pause attenuator to yield a read-through product of defined length, RNAs. (A) RNA secondary structure predicted for the E. coli trp we estimated the level of read-through at the E. coli trp pause RNA. RNA polymerase pausing occurs at the position marked attenuator to be -4% under our assay conditions. Except for by the boldface arrow. The G -+ A change that is present in trpL29 certain situations reported below, this level did not vary is indicated. Structure 1:2 is thought to be the primary pause signal between wild-type and trpL29 templates or with (10). The G shown in parentheses is the next nucleotide to be added significantly upon resumption of transcription by the paused transcription com- changes in conditions. In general, lower levels ofleader RNA plex. The next 48 nucleotides (bases 93-140, not shown) form a (caused by increased read-through) would be expected under termination structure (3:4). An alternative RNA structure (not conditions favoring decreased pausing (for example, higher shown), in which RNA segment 2 is paired with RNA segment 3, is GTP concentrations; ref. 10). A lower leader RNA level thought to form when a ribosome stalls at the tryptophan codons; this associated with a decrease in pausing would elevate the structure, the antiterminator, causes transcription read-through at relative level of pause RNA. Therefore, changes in read- the attenuator by preventing the formation of structure 3:4, the through at the attenuator would minimize rather than exag- terminator. (B) RNA secondary structures predicted for the S. gerate the effects reported here. In all of our experiments, marcescens trp leader transcript. RNA structures 1:2, 2:3 (not comparison of pause RNA levels to the level of RNA I (Fig. shown), and 3:4 (not shown) are believed to function as described for A. The boldface arrow marks the 3' end of the pause transcript. As 1) gave results very similar to the comparisons of pause RNA for the wild-type E. coli template, the G (in parentheses) is the next to leader RNA that we report. nucleotide added to the 3' end of S. marcescens pause RNA. The RNA segments removed by the various deletions used are indicated. RESULTS Translation Relieves Pausing During Coupled Transcrip- plates containing mutations that should limit leader peptide tion/Translation of the tp Leader Region. We wished to synthesis. The E. coli mutation trpL29 replaced the leader examine the effect of translation of the trp leader peptide peptide initiation codon AUG by AUA (Fig. 1A). In vivo this coding region on the rate of release of the paused transcrip- change is thought to limit or prevent leader peptide synthesis tion complex that forms during transcription of the leader because it lowers trp operon expression and reduces the region in a coupled system. To enhance RNA polymerase read-through at the attenuator normally associated with pausing we did not add carrier GTP to the labeled GTP; thus, tryptophan starvation (24). The four S. marcescens trp leader most of the GTP present was provided by the cell-free deletions used, AtrpL213, AtrpL221, AtrpL307, and extract. We estimated the concentration of GTP as 10 ,uM by AtrpL310, remove the leader peptide start codon and ribo- comparing 32p incorporation when [a-32P]UTP of known some binding site; these deletions also reduce trp operon specific radioactivity, or labeled GTP, was added. At the expression in vivo (Fig. 1B) (17, 18) and eliminate the normal GTP concentration employed, pausing was readily detected tryptophan starvation response. with the E. coli wild-type template (Figs. 1 and 2). At As shown in Fig. 2A, the pause RNA complex formed on concentrations of GTP as high as 200 ,uM, paused complex the trpL29 template is longer lived than the paused complex formation was still detectable (data not shown); however, formed on the wild-type E. coli template. Calculations based transcript elongation and paused complex release were too on these and many similar experiments indicate that the rate rapid to permit quantitation of relative pause RNA levels. of release of the trpL29 paused complex is slower by a factor To determine iftranslation ofthe leader transcript released of 2-2.5 than release of the wild-type complex. Though this the paused transcription complex, we compared pausing in difference is small, it is reproducible. In contrast, when the the coupled system on wild-type DNA templates and tem- S. marcescens Atrp221 template is used the paused transcrip- Downloaded by guest on October 2, 2021 Biochemistry: Landick et aL Proc. Natl. Acad Sci USA 82 (1985) 4665

A WT trpL29 Table 1. RNA polymerase pausing during coupled transcription/ translation with trp leader templates 0.5 1 1.5 2 2.5 0.5 1 1.5 2 2.5 min Relative pause RNA 7aAwm.-~- Paused complex level* (coupled system) half-life, sec No Template (purified system) additive Kg Cm E. coli wild type 55 7.4 33 26 trpL29 55 16 43 31 S. marcescens wild type 41 5.4 25 13 AtrpL2l3 36 56 54 60 AtrpL221 30 50 43 49 AtrpL307 55 66 68 65 AtrpL3l0 49 41 44 45 1". AL L The paused complex half-lives in purified and coupled transcrip- tion reaction mixtures were determined. Kasugamycin (Kg) (200 gg/ml) and chloramphenicol (Cm) (80 Ag/ml) were added to the RNAI cell-free extract prior to mixing with the other reaction components. *Relative pause RNA level is the molar percent of pause RNA p relative to the total molar amounts ofpause plus leader RNAs, after a 2-min labeling period. M- tion complex is released more slowly by a factor ofalmost 10 than the paused complex formed on the corresponding wild-type template (Fig. 2B, Table 1). Table 1 summarizes pausing data obtained with these four templates and with B WT A221 templates containing the S. marcescens deletions AtrpL213, A B C D E A R n n r AtrpL307, and AtrpL3J0. All ofthe deletion mutant templates produce very long-lived paused transcription complexes. To control for differences in pausing not due to translation, we measured paused complex half-lives with the same templates in a purified transcription system (Table 1). No large differences were observed between the wild-type and mutant templates. To determine whether the trpL29 pause and leader RNAs were as susceptible to degradation as wild-type transcripts in the S-30 system, we measured the degradation of purified, labeled wild-type and trpL29 RNAs L added to the S-30 reaction mixture. We found that the half-life for both leader RNAs was about 10 min, whereas both pause RNA species disappeared with a half-life of -20 sec. Deg- _- ~L radation therefore cannot account for the observed increase P. in the relative level of pause RNA caused by the trpL29 RNAI _ **t s RNAI mutation. In other experiments we verified that there was ..,w,_ sufficient GTP present (>7 uM throughout incubation) for transcription and translation to proceed. P Translation Inhibitors Increase Paused Complex Half-Life on Wild-Type Templates in the Coupled System. To provide additional evidence that the increased paused complex sta- bility observed with mutant templates was caused by the absence of translation, we examined the effect of the trans- lation inhibitors kasugamycin and chloramphenicol. We ex- FIG. 2. Pause RNAs detected during coupled transcription/ that these antibiotics would increase the relative translation reactions. (A) E. coli wild-type (WT) and trpL29 pause pected RNA release after addition of rifampicin to coupled tran- pause RNA levels observed only with the wild-type tem- scription/translation reactions. Rifampicin was added at 10 ,ug/ml plates. The addition of200 ,g ofkasugamycin per ml reduced after a 2-min labeling period and aliquots were added to phenol at the [35S]methionine incorporation by 50% (data not shown) and times indicated. RNA I is the 106- to 109-nucleotide transcript from caused a 4.5-fold increase in the relative amount of pause the colEl replication origin of the plasmid templates (25); L, leader RNA with both the E. coli and S. marcescens wild-type RNA; P, pause RNA. All RNA species were identified by RNase T1 templates (Table 1). Kasugamycin did not affect pausing on oligonucleotide fingerprinting (data not shown). (B) Pausing on S. marcescens wild-type (WT) and AtrpL221 templates in the presence the S. marcescens deletion templates, whereas it increased and absence of kasugamycin. Kasugamycin was added to the pausing on the trpL29 template -2.5-fold. The addition of 80 cell-free extract prior to mixing with the other reaction components. ,ug ofchloramphenicol per ml reduced [35S~methionine incor- The RNAs were labeled for 2 min at 370C with [a-32P]GTP after a poration by 95% and increased pausing on the wild-type E. 15-min preincubation. Chase reactions (lanes A and E) received 550 coli template 3.5-fold, on the wild-type S. marcescens tem- /AM unlabeled GTP at the end of the labeling period and were incubated for an additional 2 min before addition ofphenol. Lanes A, chase of lanes B; lanes B, no kasugamycin; lanes C, 80 ,ug of to A. Pause and leader RNAs were identified by comigration with kasugamycin per ml; lanes D, 200 ,ug of kasugamycin per ml; lanes authentic samples synthesized in a purified transcription system E, chase of lanes D. RNA I, L, and P are defined in the legend (data not shown). Downloaded by guest on October 2, 2021 4666 Biochemistry: Landick et aL Proc. NatL Acad ScL USA 82 (1985)

plate 2.5-fold, and on the trpL29 template 2-fold but had no Table 2. Pausing in extracts of a tryptophanyl-tRNA effect on the deletion templates (Table 1). If we postulate that synthetase mutant limited translation of trpL29 RNA is possible (see Discus- Wild-type sion), then we may conclude that addition of translation tryptophanyl-tRNA Relative pause inhibitors increased paused complex stability only when synthetase % read-through RNA level translation of the leader peptide coding region was possible. Pausing in S-30 Extracts Prepared from RNA Polymerase 40 8.5 Mutants. We also examined pausing and paused complex + 5 10 release by using S-30 extracts from E. coli strains containing Fourteen and one-half micrograms of purified wild-type trypto- mutations in rpoB, the structural gene for the 8 subunit of phanyl-tRNA synthetase was added to a 300- l reaction mixture RNA polymerase. Polymerases from these strains either prior to the 15-min preincubation period. After 2 min oflabeling with decrease (rpoB2) or increase (rpoB7 and rpoB8) pausing in a [a-32P]GTP, the reaction was stopped by addition of phenol. The E. purified transcription system (8, 26). We determined the coli wild-type plasmid template used in this experiment did not allow relative pause RNA level by using the seven templates listed direct measurement of percent read-through. Therefore, percent read-through was determined by comparing the ratio of leader RNA in Table 1 (data not shown). We observed that rpoB7 and to RNA I (see legend to Fig. 2) with this same ratio determined in rpoB8 increased pausing only on translation-defective tem- other experiments in which the levels of pause RNA, RNA I, leader plates, whereas rpoB2 decreased pausing on all templates. RNA, and read-through RNA all could be measured. The relative These results suggest that the translating ribosome can pause RNA level without added tRNA synthetase has been corrected overcome the effect of a mutant polymerase that increases for the lower leader RNA level caused by 40%o read-through. Direct pausing and that ribosome movement is the rate-determining comparison of the pause RNA level to the RNA I level gave results step that normally controls pause complex release. similar to those reported here. Stalled at the Tryptophan Codons of the Leader Transcript Release the Paused Transcription Complex. Since In this report we have demonstrated that mutations and the tandem tryptophan codons in the trp leader transcript are inhibitors that prevent synthesis of the leader peptide in a located in the 5' strand of the pause RNA hairpin, a coupled transcription/translation system increase the level of translating ribosome stalled at the tryptophan codons should pause RNA. Presumably this increase results from an in- release the paused complex. To test this possibility, we crease in the half-life for paused complex release. The effect performed transcription pausing assays (Table 2) using an of translation on pausing was most clearly demonstrated by S-30 extract prepared from a mutant strain (trpS9969) con- comparing the relative levels of pause RNA produced by the taining a temperature-sensitive tryptophanyl-tRNA synthe- wild-type S. marcescens template with those produced by tase (20, 27). This extract gave :50% read-through at the trp deletion templates lacking the leader peptide ribosome bind- attenuator, indicating that an appreciable fraction of ing site segment (Table 1). A relative pause RNA level that ribosomes synthesizing the leader peptide stall at the was lower by a factor of 10 was obtained with the wild-type tryptophan codons (20). After correction for the different template. We believe that this reduction results from ribo- amounts ofleader RNA, or in comparison to the level ofRNA some disruption of RNA secondary structure 1:2 (Fig. 1 A I, the relative pause RNA level observed in the trpS9969 and B), the structure that is thought to serve as the RNA extract was similar with and without addition of purified polymerase pause signal (6-11). With mutant RNA wild-type tryptophanyl-tRNA synthetase. This suggests that polymerases that increase pausing in a purified system the ribosomes that stall at the tandem tryptophan codons release effect of translation on pause release is even greater since the paused transcription complex. normal pause RNA levels are observed with translation- competent templates, whereas pausing is prolonged on DISCUSSION translation-defective templates. Addition of kasugamycin or chloramphenicol to the S-30 A key postulate of the transcription attenuation model for system increased the level of pause RNA, as would be amino acid biosynthetic operons is that synthesis of the expected ifthese compounds inhibited translation. However, leader peptide is closely coupled to transcription ofthe leader neither had as great an effect as the leader deletions, and region. If the ribosome synthesizing the leader peptide is to kasugamycin was more effective than chloramphenicol. determine which alternative transcript secondary structure These findings are consistent with the different mechanisms forms, transcription of the distal portion of the trp leader ofaction ofthese antibiotics (30) and their behavior in in vitro region must be synchronized with translation of the leader systems (31). peptide coding region. Although many aspects of the atten- One question raised by our results is why the trpL29 uation model have received considerable experimental sup- mutation has only a 2-fold effect on the relative level ofpause port (1-14, 17-21, 23, 24, 28, 29), the required coupling of RNA. Together with the fact that both chloramphenicol and transcription and translation has not been demonstrated kasugamycin increase the pause RNA level obtained with the directly. Studies in vivo have shown that transcription ter- trpL29 template, this observation suggests that there may be mination at the trp attenuator can be relieved completely by some translation initiation at the trpL29 AUA codon in the tryptophan starvation (17). According to the attenuation coupled system. In vivo, in the presence of excess model this observation implies that a ribosome has stalled at tryptophan, the trpL29 mutation causes a reduction by a the tryptophan codons of the leader peptide coding region of factor of 4 in trp operon expression. In tryptophan-starved each trp transcript as the trp leader region was being cultures the trpL29 mutation limits trp operon expression to transcribed. The discovery of transcription pausing at posi- 10-20% that ofwild type. These findings have been explained tion 92 of the trp leader transcript suggested a possible by postulating that the trpL29 change significantly reduces coupling mechanism (6, 11). If the paused transcription leader peptide synthesis. It is evident that the effects we have complex forms in vivo during transcription of the trp leader observed in vitro are not as great as would be expected from region, and if it is released as the ribosome synthesizes the the in vivo behavior of strains with the trpL29 mutation. It is leader peptide, then transcription and translation would be conceivable that the trpL29 AUA codon is translated more effectively coupled. If this is the mechanism of coupling, the efficiently in vitro than in vivo or that inefficient translation half-life for pause RNA complex release should increase ofthe leader peptide codingregion has a more profound effect when translation of the leader peptide coding region is in the growing cell. Since the rate of transcription is much prevented. slower in the coupled system, it is possible that translation Downloaded by guest on October 2, 2021 Biochemistry: Landick et aL Proc. NatL AcadJ Sci USA 82 (1985) 4667 initiation at the trpL29 AUA codon, although inefficient, ed by grants from the U.S. Public Health Service (GM-09738) and the occurs rapidly enough in vitro to release most of the paused American Heart Association (69-015). R.L. and J.C. are U.S. Public transcription complexes. Health Service Postdoctoral Fellows supported by Awards GM- Direct measurement of the half-life of paused trp leader 09151 and GM-09427, respectively. C.Y. is a Career Investigator of transcription complexes in the S-30 system would have the American Heart Association. allowed a more direct analysis of the question posed in this 1. Bauer, C. E., Carey, J., Kasper, L., Lynn, S., Waechter, D. & paper. Although such measurements have been possible in Gardner, J. (1983) in Prokaryotic Gene Expression, eds. purified transcription systems by synchronizing initiation (6, Beckwith, J., Davies, J. & Gallant, J. (Cold Spring Harbor 9, 10), we were unable to devise methods to allow synchro- Laboratory, Cold Spring Harbor, NY), pp. 65-89. nous transcription initiation in our coupled cell-free system. 2. Kolter, R. L. & Yanofsky, C. (1982) Annu. Rev. Genet. 16, Furthermore, by labeling the wild-type E. coli trp leader 113-134. transcript with short pulses of [a-32P]GTP after addition of 3. Yanofsky, C. (1981) Nature (London) 289, 751-758. rifampicin, we found that active RNA polymerase molecules 4. Platt, T. (1981) Cell 24, 10-23. 5. Yanofsky, C. (1984) Mol. Biol. Evol. 1, 143-161. located at the and at sites upstream ofthe pause site 6. Winkler, M. E. & Yanofsky, C. (1981) Biochemistry 20, synthesized substantial amounts ofpause RNA 2-3 min after 3738-3744. addition of rifampicin. Thus, the apparent 30- to 40-sec 7. Farnham, P. J. & Platt, T. (1981) Nucleic Acids Res. 9, half-life that can be estimated from our uncorrected data is a 563-577. gross overestimate of the true half-life for wild-type E. coli 8. Fisher, R. & Yanofsky, C. (1983) J. Biol. Chem. 258, paused complex release. From these and other experiments 8146-8150. we have calculated that under our conditions polymerase 9. Fisher, R. & Yanofsky, C. (1983) J. Biol. Chem. 258, molecules required %0.8 sec to add each guanine residue 9208-9212. during elongation and that release ofpaused RNA complexes 10. Landick, R. & Yanofsky, C. (1984) J. Biol. Chem. 259, 11550-11555. actually occurred with a half-life close to 5-7 sec. Given the 11. Fisher, R., Das, A., Kolter, R., Winkler, M. E. & Yanofsky, complex kinetics necessary to describe these multiple events, C. (1985) J. Mol. Biol. 182, 397-409. we found that measuring relative pause RNA levels allowed 12. Hauser, C. A., Sharp, J. A., Hatfield, L. K. & Hatfield, a more convenient comparison ofthe rate ofrelease ofpaused G. W. (1985) J. Biol. Chem. 260, 1765-1770. transcription complexes on different templates and under 13. Gardner, J. F. (1982) J. Biol. Chem. 257, 3896-3904. different reaction conditions. 14. Farnham, P., Greenblatt, J. & Platt, T. (1982) Cell 27, 523-531. The results reported here strongly support a role for 15. Birnboim, H. & Doly, J. (1979) Nucleic Acids Res. 7, ribosome release of the paused transcription complex in the 1513-1523. coupling of transcription and translation (32) in the trp leader 16. Silhavy, T. J., Berman, M. Z. & Enquist, L. W., eds. (1984) Experiments with Gene Fusions (Cold Spring Harbor Labora- region. We recognize that the rates of ribosome and RNA tory, Cold Spring Harbor, NY). polymerase movement in the cell-free system we have 17. Stroynowski, I. & Yanofsky, C. (1982) Nature (London) 298, employed are substantially slower than the rates occurring in 34-38. vivo. Hence, our findings do not constitute proof that 18. Stroynowski, I., Kuroda, M. & Yanofsky, C. (1983) Proc. ribosome release of the-paused transcription complex does Natl. Acad. Sci. USA 80, 4833-4837. synchronize transcription and translation during attenuation. 19. Kolter, R. & Yanofsky, C. (1984) J. Mol. Biol. 175, 299-312. Clearly, however, translation of the leader peptide coding 20. Das, A., Crawford, I. P. & Yanofsky, C. (1982) J. Biol. Chem. region can release paused transcription complexes formed in 257, 8795-8798. a cell-free system. A definitive demonstration of the physi- 21. Zalkin, H., Yanofsky, C. & Squires, C. L. (1974) J. Biol. ological role of this model will require demonstration of Chem. 249, 465-475. paused complex release in vivo. In other studies we have 22. Gesteland, R. F. (1966) J. Mol. Biol. 16, 67-84. shown that the does form in the 23. Oxender, D. L., Zurawski, G. & Yanofsky, C. (1979) Proc. paused transcription complex Natl. Acad. Sci. USA 76, 5524-5528. trp leader region in vivo (unpublished data). If ribosome 24. Zurawski, G., Elseviers, D., Stauffer, G. & Yanofsky, C. release of transcription pause complexes is used generally to (1978) Proc. Natl. Acad. Sci. USA 75, 5988-5992. facilitate the coupling of transcription and translation, paus- 25. Morita, M. & Oka, A. (1979) Eur. J. Biochem. 97, 435-443. ing may play a major physiological role by permitting the cell 26. Yanofsky, C. & Horn, V. (1981) J. Bacteriol. 145, 1334-1341. to avoid activation of Rho-mediated transcription termina- 27. Doolittle, W. F. & Yanofsky, C. (1968) J. Bacteriol. 95, tion. Thus, whenever synthesis of a messenger RNA pro- 1283-1294. ceeds without concomitant translation, as occurs when 28. Platt, T., Squires, C. & Yanofsky, C. (1976) J. Mol. Biol. 103, translation initiation is inefficient, pausing may delay tran- 411-420. script elongation until the initial translating ribosome releases 29. Zurawski, G. & Yanofsky, C. (1980) J. Mol. Biol. 142, the paused complex. Pausing thus would prevent the syn- 123-129. thesis ofthe long, ribosome-free messenger segment required 30. Petska, S. (1977) in Molecular Mechanisms ofProtein Biosyn- for Rho entry. thesis, eds. Weissbach, H. & Petska, S. (Academic, New York), pp. 467-553. We thank Vivian Berlin, Mitzi Kuroda, and Valley Stewart for 31. Julian, G. R. (1965) J. Mol. Biol. 12, 9-16. their critical reading of this manuscript. These studies were support- 32. Stent, G. S. (1966) Proc. R. Soc. London Ser. B 164, 181-197. Downloaded by guest on October 2, 2021