Promoter Clearance by T7 RNA Polymerase

Footprinting, fluorescence, and x-ray structural information from the initial, promoter-bound complex of T7 RNA polymerase describes the very beginning of the initiation of transcription, whereas recent fluorescence and biochemical studies paint a preliminary picture of an elongation complex. The current work focuses on the transition from an initially transcribing, promoter-bound complex to an elongation complex clear of the promoter. Fluorescence quenching is used to follow the melted state of the DNA bubble, and a novel approach using a locally mismatched fluorescent base analog reports on the local structure of the heteroduplex. Fluorescent base analogs placed at positions −2 and −1 of the promoter indicate that this initially melted, nontranscribed region remains melted as the polymerase translocates through to position +8. In progressing to position +9, this region of the DNA bubble begins to collapse. Probes placed at positions +1 and +2 of the template strand indicate that the 5′ end of the RNA remains in a heteroduplex as the complex translocates to position +10. Subsequent translocation leads to sequential dissociation of the first 2 bases of the RNA. These results show that the initially transcribing complex bubble can reach a size of up to 13 base pairs and a maximal heteroduplex length of 10 base pairs. They further indicate that initial bubble collapse precedes dissociation of the 5′ end of the RNA. Footprinting, fluorescence, and x-ray structural information from the initial, promoter-bound complex of T7 RNA polymerase describes the very beginning of the initiation of transcription, whereas recent fluorescence and biochemical studies paint a preliminary picture of an elongation complex. The current work focuses on the transition from an initially transcribing, promoter-bound complex to an elongation complex clear of the promoter. Fluorescence quenching is used to follow the melted state of the DNA bubble, and a novel approach using a locally mismatched fluorescent base analog reports on the local structure of the heteroduplex. Fluorescent base analogs placed at positions −2 and −1 of the promoter indicate that this initially melted, nontranscribed region remains melted as the polymerase translocates through to position +8. In progressing to position +9, this region of the DNA bubble begins to collapse. Probes placed at positions +1 and +2 of the template strand indicate that the 5′ end of the RNA remains in a heteroduplex as the complex translocates to position +10. Subsequent translocation leads to sequential dissociation of the first 2 bases of the RNA. These results show that the initially transcribing complex bubble can reach a size of up to 13 base pairs and a maximal heteroduplex length of 10 base pairs. They further indicate that initial bubble collapse precedes dissociation of the 5′ end of the RNA. initially transcribing complex base pair The recent past has seen a number of structures of RNA polymerases, ranging from RNA polymerase II (1Cramer P. Bushnell D.A. Kornberg R.D. Science. 2001; 292: 1863-1876Crossref PubMed Scopus (966) Google Scholar, 2Gnatt A.L. Cramer P. Fu J. Bushnell D.A. Kornberg R.D. Science. 2001; 292: 1876-1882Crossref PubMed Scopus (744) Google Scholar) to a smaller bacterial RNA polymerase (3Zhang G. Campbell E.A. Minakhin L. Richter C. Severinov K. Darst S.A. Cell. 1999; 98: 811-824Abstract Full Text Full Text PDF PubMed Scopus (670) Google Scholar) to the simplest well studied system: the single-subunit T7 RNA polymerase (4Sousa R. Chung Y.J. Rose J.P. Wang B.C. Nature. 1993; 364: 593-599Crossref PubMed Scopus (341) Google Scholar, 5Jeruzalmi D. Steitz T.A. EMBO J. 1998; 17: 4101-4113Crossref PubMed Scopus (152) Google Scholar, 6Cheetham G.M. Jeruzalmi D. Steitz T.A. Nature. 1999; 399: 80-83Crossref PubMed Scopus (274) Google Scholar, 7Cheetham G.M. Steitz T.A. Science. 1999; 286: 2305-2309Crossref PubMed Scopus (290) Google Scholar). Structures are available for polymerases without DNA, with DNA bound, and with either a very short initial transcript or with longer transcripts, corresponding to elongation. Lacking in any system, however, is x-ray structural information on the critical transition from an unstable initially transcribing complex (ITC)1to a stable elongation complex. This transition occurs at about 10 base pairs (bp) in all RNA polymerases, suggesting that the transition is a fundamental feature of transcription, independent of the specific system (8Ebright R.H. J. Mol. Biol. 2000; 304: 687-698Crossref PubMed Scopus (189) Google Scholar). What is the nature of the transcription bubble at this critical, early transition point? For the single-subunit RNA polymerase from T7, a crystal structure is available with a 3-base RNA transcript at the active site (7Cheetham G.M. Steitz T.A. Science. 1999; 286: 2305-2309Crossref PubMed Scopus (290) Google Scholar). Modeling from that structure, Cheetham and Steitz (7Cheetham G.M. Steitz T.A. Science. 1999; 286: 2305-2309Crossref PubMed Scopus (290) Google Scholar) predicted that as the polymerase transcribes forward, the enzyme could accommodate no more than a 3-base heteroduplex. Recent results have shown, however, that in a stably elongating complex stalled clear of the promoter, near position +15, the transcription bubble extends about 8–9 bp upstream of the stall site, consistent with a heteroduplex size of ∼8 bp (9Huang J. Sousa R. J. Mol. Biol. 2000; 303: 347-358Crossref PubMed Scopus (68) Google Scholar, 10Temiakov D. Mentesana P.E. Ma K. Mustaev A. Borukhov S. McAllister W.T. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 14109-14114Crossref PubMed Scopus (74) Google Scholar, 11Liu C. Martin C.T. J. Mol. Biol. 2001; 308: 465-475Crossref PubMed Scopus (146) Google Scholar). Cross-linking studies report that in a complex stalled at position +23, RNA at position −9 relative to the stall site cross-links to RNA polymerase but not to DNA, whereas RNA at positions closer to the last incorporated base cross-links to DNA, again consistent with an 8-bp heteroduplex (10Temiakov D. Mentesana P.E. Ma K. Mustaev A. Borukhov S. McAllister W.T. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 14109-14114Crossref PubMed Scopus (74) Google Scholar). Footprinting studies have indicated that complexes stalled at position +6 retain promoter occupancy, as in the crystal structure, but complexes stalled at position +15 have cleared the promoter (9Huang J. Sousa R. J. Mol. Biol. 2000; 303: 347-358Crossref PubMed Scopus (68) Google Scholar, 12Ikeda R.A. Richardson C.C. Proc. Natl. Acad. Sci. U. S. A. 1986; 83: 3614-3618Crossref PubMed Scopus (147) Google Scholar). It is reasonable to expect a substantial difference in these structures; perhaps the initially transcribing complex has different structural constraints. We have previously used fluorescent base analogs to map DNA melting in both the initial promoter-bound complex and in a stably elongating complex (beyond the transition) in the model enzyme T7 RNA polymerase (11Liu C. Martin C.T. J. Mol. Biol. 2001; 308: 465-475Crossref PubMed Scopus (146) Google Scholar, 13Újvári A. Martin C.T. Biochemistry. 1996; 35: 14574-14582Crossref PubMed Scopus (94) Google Scholar). The details of the model for the initial melted complex derived from the early fluorescence study have been confirmed by recent crystal structures (6Cheetham G.M. Jeruzalmi D. Steitz T.A. Nature. 1999; 399: 80-83Crossref PubMed Scopus (274) Google Scholar), and the results from the latter are consistent with biochemical probes of a similarly stalled elongation complex (9Huang J. Sousa R. J. Mol. Biol. 2000; 303: 347-358Crossref PubMed Scopus (68) Google Scholar). In the current work, we turn this approach toward mapping the collapse of the initially melted, untranscribed region of the DNA and the initial peeling away of the RNA from the heteroduplex as the ITC progresses through and beyond position +10. The new finding that pyrrolo-dC in a mismatch context has typical fluorescence values higher than that of the same probe in single-stranded DNA (which, in turn, has higher fluorescence than in double-stranded DNA), coupled with the fact that fluorescence is quenched in a heteroduplex, now allows us to probe not only the melted state of the DNA but also for the presence of a heteroduplex. The results present a detailed structural picture of promoter clearance in the T7 RNA polymerase model system. T7 RNA polymerase was prepared fromEscherichia coli strain BL21, carrying the overproducing plasmid pAR1219 (kindly supplied by F. W. Studier), in which RNA polymerase is expressed under inducible control of the lacUV5 promoter (14Davanloo P. Rosenberg A.H. Dunn J.J. Studier F.W. Proc. Natl. Acad. Sci. U. S. A. 1984; 81: 2035-2039Crossref PubMed Scopus (729) Google Scholar). The enzyme was purified, and the concentration was determined (ε280 = 1.4 × 105m/cm) as described previously (15King G.C. Martin C.T. Pham T.T. Coleman J.E. Biochemistry. 1986; 25: 36-40Crossref PubMed Scopus (90) Google Scholar). Purity of the enzyme (>95%) was verified by SDS-denaturing polyacrylamide gel electrophoresis. Oligonucleotides were synthesized by the phosphoramidite method on an Expedite 8909 DNA/RNA synthesizer. The standard phosphoramidites were purchased from CPG Inc. The fluorescent cytosine analog pyrrolo-dC was incorporated into oligonucleotides using furano-dT phosphoramidite (Glen Research Corp.) as described previously (11Liu C. Martin C.T. J. Mol. Biol. 2001; 308: 465-475Crossref PubMed Scopus (146) Google Scholar). Single strands were purified using an Amberchrom CG-16C reverse phase resin as described (16Schick C. Martin C.T. Biochemistry. 1993; 32: 4275-4280Crossref PubMed Scopus (25) Google Scholar). Purity of the single-stranded oligonucleotides was confirmed by denaturing (urea) gel electrophoresis of 5′ end-labeled single strands. Double-stranded templates were prepared by heating a 1:1 mixture of complementary single strands in Tris-EDTA buffer (10 mm Tris, pH 7.8, 1 mm EDTA) at 75 °C for 5 min. The samples were then allowed to cool slowly to room temperature. The constructs were either used immediately or stored at −20 °C. Fluorescence measurements were carried out with a Photon Technology International L-format fluorimeter with a 75-watt arc lamp and both emission and excitation monochrometers using a 75-μl (light path, 3 × 3 mm; center, 15 mm) ultramicro cell (Hellma). Fluorescence emission from pyrrolo-dC was detected at 460 nm, with excitation at 350 nm; slits on both channels were set to 5 nm. Fluorescence from 2-aminopurine was obtained similarly but with excitation at 315 nm and emission detected at 370 nm. For measurements of fluorescence from 2-aminopurine, fluorescence is reported after subtraction of the enzyme background. All fluorescence experiments were carried out in fluorescence buffer: 30 mm HEPES, pH 7.8, 15 mm magnesium acetate, 25 mm potassium glutamate, 0.25 mm EDTA, and 0.05% Tween 20 (Calbiochem; 10% protein grade) in a sample compartment thermostated at 25 °C. The fluorescence changes recorded in all experiments represent an average of three measurements. In the chase experiments, the fluorescence from double-stranded DNA (1.0 μm) was first recorded alone and then after the addition of each of the following: enzyme (to ∼1.0 μm final concentration), GTP and ATP, 3′-dCTP (or CTP), and UTP (to 1000 μm each), with each addition occurring at 3-min intervals. For samples with two fluorophores, the same protocol was followed, except that during the measurement, the monochrometers were switched, under computer control, alternately between the excitation and emission pairs above. Transcription reactions were carried out in the above fluorescence buffer at 25 °C in a volume of 20 μl containing nucleoside triphosphates as described below and [α-32P]GTP (specific activity, 800 Ci/mmol; PerkinElmer Life Sciences). Reactions were stopped by the addition of an equal volume of quenching solution (95% formaldehyde, 50 mmEDTA, 0.01% bromphenol blue, 0.01% xylene cyanol). The enzyme and DNA (1.0 μm each) were incubated in transcription buffer (lacking nucleoside triphosphates) for 4 min at room temperature and then divided into two equivalent fractions. GTP and ATP (including [α-32P]GTP) were added to the first fraction, which was then incubated for 3 min at room temperature and quenched. The second fraction was similarly incubated with GTP and ATP for 3 min, and then 3′-dCTP was added to a final concentration of 1000 μm. This was then incubated for 3 min and quenched. All transcription samples were heated to 95 °C for 2 min and loaded onto 20% polyacrylamide-6 m urea gels. After electrophoresis, gels were dried and quantified using a Molecular Dynamics Storm 840 PhosphorImager. In the initial formation of the open complex, base pairs at positions −4 through +3 are melted open (6Cheetham G.M. Jeruzalmi D. Steitz T.A. Nature. 1999; 399: 80-83Crossref PubMed Scopus (274) Google Scholar, 13Újvári A. Martin C.T. Biochemistry. 1996; 35: 14574-14582Crossref PubMed Scopus (94) Google Scholar). As the polymerase translocates through to position +3, the upstream edge (position −4) of this initial bubble remains open (7Cheetham G.M. Steitz T.A. Science. 1999; 286: 2305-2309Crossref PubMed Scopus (290) Google Scholar). However, translocation to position +15 results in the closure of the initial bubble and the loss of promoter contacts (11Liu C. Martin C.T. J. Mol. Biol. 2001; 308: 465-475Crossref PubMed Scopus (146) Google Scholar, 12Ikeda R.A. Richardson C.C. Proc. Natl. Acad. Sci. U. S. A. 1986; 83: 3614-3618Crossref PubMed Scopus (147) Google Scholar). In a complex stalled at position +15, the heteroduplex extends ∼8 bp upstream of the last base incorporated (9Huang J. Sousa R. J. Mol. Biol. 2000; 303: 347-358Crossref PubMed Scopus (68) Google Scholar, 10Temiakov D. Mentesana P.E. Ma K. Mustaev A. Borukhov S. McAllister W.T. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 14109-14114Crossref PubMed Scopus (74) Google Scholar, 11Liu C. Martin C.T. J. Mol. Biol. 2001; 308: 465-475Crossref PubMed Scopus (146) Google Scholar). Between these two states, the RNA polymerase must clear the promoter and become a stable elongation complex. At what point between positions +3 and +15 does the initial bubble collapse? At what point does the RNA begin to peel away from the heteroduplex? Fluorescence from base analogs placed site-specifically within the DNA reports on the melted state near the analog. To monitor fluorescence changes at specific positions as the RNA polymerase translocates along the template DNA, DNA constructs (summarized in TableI) were prepared, which allow walking of the RNA polymerase to positions +7–12 in the presence of only GTP and ATP as substrates. Subsequent addition of 3′-dCTP allows further translocation by only 1 base pair, to positions +8–13, ensuring that misincorporation does not lead to heterogeneity in the stalled complexes. In some experiments below, subsequent addition of UTP allows binding of the next substrate nucleotide (and translocation by 1 more base), but without covalent bond formation.Table ISequences of DNA constructs used throughout this studyNomenclatureDNA sequence downstream of position −52AP−2TGA (7)…TATAGGGAGAGCTTACGGTTATCTAGATC…ATATCCCTCTCGAATGCCAATAGATCTAG2AP−2TGA (8)…TATAGGGAGAGGCTTCGGTTATCTAGATC…ATATCCCTCTCCGAAGCCAATAGATCTAG2AP−2TGA (9)…TATAGGGAGAGAGCTTCGTTATCTAGATC…ATATCCCTCTCTCGAAGCAATAGATCTAG2AP−2TGA (10)…TATAGGGAGAGGAGCTTCTTATCTAGATC…ATATCCCTCTCCTCGAAGAATAGATCTAGpyC+1TGA (10)…TATAGGGAGAGGAGCTTCTATCCTAGATC…ATATCCCTCTCCTCGAAGATAGGATCTAGpyC+1TGA (11)…TATAGGGAGAAGGAGCTTCTATCTAGATC…ATATCCCTCTTCCTCGAAGATAGATCTAGpyC+2TGA (10)…TATAGGGAGAGGAGCTTCTATCCTAGATC…ATATCCCTCTCCTCGAAGATAGGATCTAGpyC+2TGA (11)…TATAGGGAGAAGGAGCTTCTATCTAGATC…ATATCCCTCTTCCTCGAAGATAGATCTAGm-pyC+1TGA (10)…TATAAGGAGAGGAGCTTCTATCCTAGATC…ATATCCCTCTCCTCGAAGATAGGATCTAGm-pyC+1TGA (11)…TATAAGGAGAAGGAGCTTCTATCTAGATC…ATATCCCTCTTCCTCGAAGATAGATCTAGm-pyC+1TGA (12)…TATAAGGAGAGAGGAGCTTCTACTAGATC…ATATCCCTCTCTCCTCGAAGATGATCTAGm-pyC+2TGA (10)…TATAGAGAGAGGAGCTTCTATCCTAGATC…ATATCCCTCTCCTCGAAGATAGGATCTAGm-pyC+2TGA (11)…TATAGAGAGAAGGAGCTTCTATCTAGATC…ATATCCCTCTTCCTCGAAGATAGATCTAGm-pyC+2TGA (12)…TATAGAGAGAGAGGAGCTTCTACTAGATC…ATATCCCTCTCTCCTCGAAGATGATCTAGpyC+11NGA (10)…TATAGGAAGGGAGGCGGAGTTTACTCATC…ATATCCTTCCCTCCGCCTCAAATGAGTAGpyC+13TGA (10)…TATAGGAAGGGAGGCGGAGTTTACTCATC…ATATCCTTCCCTCCGCCTCAAATGAGTAG2AP−2TpyC+8NGA (7)…TATAGGGAGAGCTTACGGTTATCTAGATC…ATATCCCTCTCGAATGCCAATAGATCTAG2AP−2TpyC+9NGA (8)…TATAGGGAGAGGCTTCGGTTATCTAGATC…ATATCCCTCTCCGAAGCCAATAGATCTAG2AP−2TpyC+10NGA (9)…TATAGGGAGAGAGCTTCGTTATCTAGATC…ATATCCCTCTCTCGAAGCAATAGATCTAGAll constructs contained the duplex promoter (nontemplate) consensus sequence TAATACGACTCAC from positions −17 through −5. The remainder of the downstream sequence for each is shown above. In the nomenclature presented, the first component describes the fluorescent base analog: pyC denotes pyrrolo-dC, and 2AP denotes (deoxy-)2-aminopurine. The subscript following denotes the position and strand, template (T) or nontemplate (N), of the analog. The number in parentheses following GA represents the predicted length of the stalled RNA in the presence of only GTP and ATP. The last three entries contain two fluorophores per duplex. Finally, m-pyC refers to constructs with A mispaired opposite pyC. In the sequences presented in the second column, the positions of the base analog fluorophores are represented in bold (A for 2AP and C for pyC). Open table in a new tab All constructs contained the duplex promoter (nontemplate) consensus sequence TAATACGACTCAC from positions −17 through −5. The remainder of the downstream sequence for each is shown above. In the nomenclature presented, the first component describes the fluorescent base analog: pyC denotes pyrrolo-dC, and 2AP denotes (deoxy-)2-aminopurine. The subscript following denotes the position and strand, template (T) or nontemplate (N), of the analog. The number in parentheses following GA represents the predicted length of the stalled RNA in the presence of only GTP and ATP. The last three entries contain two fluorophores per duplex. Finally, m-pyC refers to constructs with A mispaired opposite pyC. In the sequences presented in the second column, the positions of the base analog fluorophores are represented in bold (A for 2AP and C for pyC). The DNA sequence TATA at positions −4 through −1 of the promoter is melted in the initial open complex (6Cheetham G.M. Jeruzalmi D. Steitz T.A. Nature. 1999; 399: 80-83Crossref PubMed Scopus (274) Google Scholar, 7Cheetham G.M. Steitz T.A. Science. 1999; 286: 2305-2309Crossref PubMed Scopus (290) Google Scholar, 13Újvári A. Martin C.T. Biochemistry. 1996; 35: 14574-14582Crossref PubMed Scopus (94) Google Scholar). Placement of 2-aminopurine at position −2 of the template strand provides a probe of the melted state of this region, as shown in earlier studies (13Újvári A. Martin C.T. Biochemistry. 1996; 35: 14574-14582Crossref PubMed Scopus (94) Google Scholar, 17Jia Y. Kumar A. Patel S. J. Biol. Chem. 1996; 271: 30451-30458Abstract Full Text Full Text PDF PubMed Scopus (104) Google Scholar, 18Bandwar R.P. Patel S.S. J. Biol. Chem. 2001; 276: 14075-14082Abstract Full Text Full Text PDF PubMed Scopus (66) Google Scholar). A set of data showing closing of the bubble is presented in Fig.1. Different DNA constructs were prepared, which in the presence of GTP and ATP allow the steady-state accumulation of complexes stalled at various positions (indicated by the upper member of each pair of numbers along the y axis). Subsequent addition of 3′dCTP allows translocation by one nucleotide (as indicated by the lower number of each pair). The results show clearly that in complexes translocated to position +7 or +8, the initially melted region (positions −2 and −1) remains melted, as evidenced by the high fluorescence relative to the double-stranded control. In complexes stalled at position +9, on each of two different constructs, the fluorescence is intermediate between that of double- and single-stranded DNA. Finally, on translocation to position +10, the fluorescence is close to that of the fully double-stranded control. The latter result shows that by position +10, the TATA region, as represented by probes at positions −2 and −1, has collapsed. In contrast, on translocation up to position +8, the initially melted region remains melted, consistent with the retention of promoter contacts. Looking more carefully at the transition itself, these data are consistent with one of three, more detailed, models. In one model, on translocation to position +9, the complexes are homogeneous, but the stacking interactions at position −2 have been altered to yield an intermediate level of fluorescence. In a clearly distinct model, on translocation to position +9, the initially melted region collapses fully in ∼40–50% of the complexes, with near complete collapse occurring on translocation to position +10. A third model suggests that on stalling at position +9, there is a distribution of translocational complexes. In this most likely scenario, the complexes are distributed between those that have incorporated the base at +9 but have not moved forward and those that have moved forward to expose the elongating base in the template strand. In this scenario, collapse of the bubble would occur on translocation to position +10. A potential issue in these types of studies involves the fact that walked complexes can undergo turnover (returning to the initial promoter positioning) during the measurement, such that steady-state measurements under any given set of substrates will yield a distribution of complexes, including complexes poised for initiation (19Mentesana P.E. Chin-Bow S.T. Sousa R. McAllister W.T. J. Mol. Biol. 2000; 302: 1049-1062Crossref PubMed Scopus (46) Google Scholar). It is important to note that the time scale of the fluorescence measurement (nanoseconds) is much faster than potential translocational events, such that this approach provides an accurate measurement of any such distribution. Direct fluorescence measurements can easily be carried out within 3–4 min of addition of the substrate NTPs, limiting overall turnover (11Liu C. Martin C.T. J. Mol. Biol. 2001; 308: 465-475Crossref PubMed Scopus (146) Google Scholar). To assess the extent of turnover within this time frame, steady-state measurements of RNA synthesis were made under conditions identical to the fluorescence studies. The results, summarized in Table II, show that turnover (maximally 1 min–1 for complexes stalled at position +7) is slower than the rate of initiation, providing evidence that the expected stalled complex should be the predominant species in the fluorescence measurements. These results also demonstrate that, as expected, the turnover rate steadily decreases with increasing position of the stall. Given that the sequence context at the stall site, as well as the initial few bases of the transcript, remain constant across these constructs, these results also demonstrate that the inherent stability of the ITC increases steadily as the enzyme translocates from position +7 to +10.Table IICharacterization of turnover on templates encoding different length RNAs under conditions that mimic the fluorescence measurementsDNA constructnGAGA [Rn]GA + 3′-dC[Rn][Rn + 1]μm2AP−2TGA(7)+72.05.01.22AP−2TGA(8)+81.71.21.82AP−2TGA(9)+91.00.40.82AP−2TGA(10)+100.80.150.9As described under “Materials and Methods,” 1.0 μmpolymerase and 1.0 μm double-stranded promoter were allowed to react for 3 min at 25° C in the presence of 1000 μm GTP, 1000 μm ATP, and [32P]GTP, conditions that mimic those in the fluorescence measurement. Reactions were quenched and electrophoresed, and bands were quantified. The concentrations of RNA (R) of stall length are presented in column GA. In an identical experiment, at 3 min, 3′-dCTP was added to a final concentration of 1000 μm, and the reaction was incubated for a further 3 min. The concentrations of RNAs of length n and n + 1 are presented in the columns GA + 3′-dC. On all of these templates, abortive products beyond 3 mer were minimal. Construct nomenclature is that of Table 1. Open table in a new tab As described under “Materials and Methods,” 1.0 μmpolymerase and 1.0 μm double-stranded promoter were allowed to react for 3 min at 25° C in the presence of 1000 μm GTP, 1000 μm ATP, and [32P]GTP, conditions that mimic those in the fluorescence measurement. Reactions were quenched and electrophoresed, and bands were quantified. The concentrations of RNA (R) of stall length are presented in column GA. In an identical experiment, at 3 min, 3′-dCTP was added to a final concentration of 1000 μm, and the reaction was incubated for a further 3 min. The concentrations of RNAs of length n and n + 1 are presented in the columns GA + 3′-dC. On all of these templates, abortive products beyond 3 mer were minimal. Construct nomenclature is that of Table 1. Our previous study of an elongation complex stalled at position +15 showed that the downstream edge of the melted bubble lies very close to the last incorporated base (11Liu C. Martin C.T. J. Mol. Biol. 2001; 308: 465-475Crossref PubMed Scopus (146) Google Scholar). We now ask whether this is true in complexes that have just cleared the promoter (as evidenced by collapse of the upstream TATA bubble). The data in Fig. 1 follow fluorescence from probes near the start site, in complexes stalled 7–11 nucleotides downstream. In these same stalled complexes, we can use fluorescence from probes placed just downstream of the stall site. The results presented in Fig. 2A show that in complexes stalled at position +10, fluorescence from a probe placed at position +11 on the nontemplate strand increases slightly, indicating only minor melting at this position just 1 base downstream of the stall site. Walking these complexes forward to position +15, to place the probe fully within the expected bubble (position −5 relative to the stall site), yields an increase in fluorescence to single-stranded levels, consistent with full opening. In a complementary experiment presented in Fig. 2B, placing the label in the template strand at position +13 and walking to position +10 yields no increase in fluorescence, consistent with the retention of duplex DNA downstream. Interestingly, walking this stalled complex to position +15 also results in no increase in fluorescence. In this case, the DNA must be melted, but the probe in the template strand is now in an RNA-DNA heteroduplex. One cannot, from the above, rule out the possibility that some stalled complexes, at position +8, for example, might transiently backtrack to the promoter, thus yielding a distribution of products at steady state. In other words, the actual footprint of an individual enzyme on the DNA might be smaller than implied by the data in Fig. 1 but reflect a steady-state scenario in which some complexes are situated at the promoter with their −2 base pairs open and their +8 base pairs closed, whereas others are situated at the +8 stall site, with their −2 bases closed and their +8 bases open. One might also worry that the introduction of base analogs in two separate constructs results in different behavior for the two constructs. To address these concerns quantitatively, we prepared a series of DNA constructs in which 2-aminopurine was placed at position −2 of the template strand, whereas pyrrolo-dC was simultaneously placed immediately downstream of the stall site on the nontemplate strand. Excitation at 315 nm with observation at 370 nm yielded a signal from 2-aminopurine, whereas excitation at 350 nm with observation at 460 nm yielded a signal from pyrrolo-dC (in all cases, no evidence of energy transfer from 2-aminopurine to pyrrolo-dC was observed). As shown in Fig. 3, a steady decrease in fluorescence from 2-aminopurine placed at position −2 of the template strand is observed in translocation from position +8 to +10, as was seen in Fig. 1. The magnitude of the change relative to the controls is consistent with an overwhelming majority of the complexes transiting from fully open to fully closed. At the same time, in constructs stalled with ATP and GTP only, the fluorescence from pyrrolo-dC at a position 1 base downstream of the initial stall site is less than, but close to, single-stranded levels, consistent with melting of the DNA 1 base pair downstream of the stall site in at least 50% of the complexes. This signal increases only slightly with the addition of 3′dCTP and negligibly with subsequent addition of UTP. This level of fluorescence is likely the true measure of a fully melted (at the stall site) complex, such that close to 100% of the complexes are stalled as anticipated. Thus, under conditions designed to stall transcription at position +8, the majority of the complexes have melted DNA at positions −2 and +8, inconsistent with transient sliding of a (smaller) fixed length bubble. The changes in fluorescence intensity observed in 2-aminopurine are thought to arise from changes in the stacking interactions surrounding the fluorophore (20Nordlund T.M. Xu D. Evans K.O. Biochemistry. 1993; 32: 12090-12095Crossref PubMed Scopus (82) Google Scholar, 21Xu D.G. Nordlund T.M. Biophys. J. 2000; 78: 1042-1058Abstract Full Text Full Text PDF PubMed Scopus (73) Google Scholar, 22Rachofsky E.L. Osman R. Ross J.B. Biochemistry. 2001; 40: 946-956Crossref PubMed Scopus (310) Google Scholar). In duplex DNA, the bases are well stacked, providing for efficient quenching of fluorescence. In contrast, in single-stranded DNA, stacking is still present but is presumably less well ordered. These results have been used to probe melted regions within DNA, both for 2-aminopurine and, more recently, for pyrrolo-dC (11Liu C. Martin C.T. J. Mol. Biol. 2001; 308: 465-475Crossref PubMed Scopus (146) Google Scholar, 13Újvári A. Martin C.T. Biochemistry. 1996; 35: 14574-14582Crossref PubMed Scopus (94) Google Scholar, 17Jia Y. Kumar A. Patel S. J. Biol. Chem. 1996; 271: 30451-30458Abstract Full Text Full Text PDF PubMed Scopus (104) Google Scholar, 23Sullivan J.J. Bjornson K.P. Sowers L.C. deHaseth P.L. Biochemistry. 1997; 36: 8005-8012Crossref PubMed Scopus (32) Google Scholar). The results shown in Fig. 4 demonstrate that the fluorescence of pyrrolo-dC in an engineered mismatched setting (opposite adenine) is higher than that of the same probe in single stranded DNA. Presumably, the mismatch imposes constraints on stacking not present in the single strand. This result can be used to our advantage in assessing the state of the bubble and of the heteroduplex. Correct pairing of the probe pyrrolo-dC in the template strand opposite guanine in the original DNA duplex, should give low fluorescence when the probe is in either a DNA:DNA duplex or in an RNA:DNA heteroduplex. Fluorescence will be high only when the template strand is in a single stranded environment. Construction of an identical DNA construct, with the exception that pyrrolo-dC is mispaired opposite adenine, provides complementary data. In this case, fluorescence will be low in the (correctly paired) RNA:DNA heteroduplex, but formation of the (mispaired) DNA:DNA duplex should show fluorescence intensities higher than that of the same probe in single (or double)-stranded DNA. To determine at what point the 5′ end of the RNA initially separates from the heteroduplex and the DNA reanneals, we placed fluorescent base analogs at positions +1 and +2, either correctly paired opposite guanine or mispaired with adenine in the nontemplate strand. The results presented in Fig.5, A and C, suggest that with one possible exception, probes placed at position +1 or +2 of the template strand remain in a duplex environment as the polymerase translocates from positions +10–12. The fluorescence is close to that of the control DNA duplex and significantly lower than that of the single-stranded DNA. However, these results cannot distinguish between a DNA-DNA or an RNA-DNA duplex. The data in Fig. 5, B and D, provide this distinction. In this case, the “duplex” control contains pyrrolo-dC in a mismatched environment, opposite adenine. In complexes stalled at position +10, fluorescence from probes placed at positions +1 and +2 is significantly lower than that of the duplex control, consistent with the presence of the heteroduplex at those positions. As the complex progresses to position +11, the fluorescence from a probe at position +1 increases, whereas that from a probe at position +2 remains constant. Progression to position +12 leads to high fluorescence from both probes (although the extent varies depending on how the complexes are walked to position +12, again suggesting some distribution of translocational complexes), consistent with the reannealing of the duplex, to provide a mismatched environment at the fluorophore. All RNA polymerases show a transition from an abortive cycling phase to a more stable elongation phase at ∼10 base pairs, a region midway between the initial promoter-bound complex and an elongation complex stalled clear of the promoter (24Uptain S.M. Kane C.M. Chamberlin M.J. Annu. Rev. Biochem. 1997; 66: 117-172Crossref PubMed Scopus (397) Google Scholar). Understanding this transition is critical to understanding both the fundamental machinery of transcription and the various regulatory processes that exploit this phenomenon. That the transition occurs at about the same position in both the multisubunit eukaryotic and prokaryotic polymerases and in the much smaller single-subunit RNA polymerase from T7 suggests a fundamental mechanism inherent to transcription. Fluorescent base analogs provide a relatively nonperturbing probe of local structural changes in the DNA, which accompany protein binding (21Xu D.G. Nordlund T.M. Biophys. J. 2000; 78: 1042-1058Abstract Full Text Full Text PDF PubMed Scopus (73) Google Scholar, 22Rachofsky E.L. Osman R. Ross J.B. Biochemistry. 2001; 40: 946-956Crossref PubMed Scopus (310) Google Scholar, 25Bierzynski A. Kozlowska H. Wierzchowski K.L. Biophys. Chem. 1977; 6: 223-229Crossref PubMed Scopus (16) Google Scholar, 26Xu D. Evans K. Nordlund T. Biochemistry. 1994; 33: 9592-9599Crossref PubMed Scopus (123) Google Scholar, 27Frey M. Sowers L. Millar D. Benkovic S. Biochemistry. 1995; 34: 9185-9192Crossref PubMed Scopus (150) Google Scholar). In particular, 2-aminopurine has been used to map melted regions of the DNA in initial promoter-bound complexes of T7 RNA polymerase (13Újvári A. Martin C.T. Biochemistry. 1996; 35: 14574-14582Crossref PubMed Scopus (94) Google Scholar, 18Bandwar R.P. Patel S.S. J. Biol. Chem. 2001; 276: 14075-14082Abstract Full Text Full Text PDF PubMed Scopus (66) Google Scholar). More recently, we have used two new fluorescent probes 2C. Liu, M. E. Hawkins, and C. T. Martin, unpublished results. to expand the repertoire of sequences that can be studied to map the melted bubble in elongation complexes stalled at position +15, clear of the promoter (11Liu C. Martin C.T. J. Mol. Biol. 2001; 308: 465-475Crossref PubMed Scopus (146) Google Scholar). The results of these studies are consistent with recent crystal structures and with biochemical studies, confirming the general utility of this approach (6Cheetham G.M. Jeruzalmi D. Steitz T.A. Nature. 1999; 399: 80-83Crossref PubMed Scopus (274) Google Scholar, 7Cheetham G.M. Steitz T.A. Science. 1999; 286: 2305-2309Crossref PubMed Scopus (290) Google Scholar, 9Huang J. Sousa R. J. Mol. Biol. 2000; 303: 347-358Crossref PubMed Scopus (68) Google Scholar, 10Temiakov D. Mentesana P.E. Ma K. Mustaev A. Borukhov S. McAllister W.T. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 14109-14114Crossref PubMed Scopus (74) Google Scholar). The quenching of the fluorescence of base analogs in DNA is thought to arise primarily from neighbor stacking interactions (22Rachofsky E.L. Osman R. Ross J.B. Biochemistry. 2001; 40: 946-956Crossref PubMed Scopus (310) Google Scholar). Such interactions are substantial in single-stranded DNA but increase significantly in the more ordered environment of a DNA duplex. The results presented here show that quenching is similarly higher in an RNA-DNA heteroduplex relative to the control single-stranded DNA. This is useful for probing template strand involvement in a hybrid but presents a problem in that one cannot readily distinguish an RNA-DNA duplex from a DNA-DNA duplex. The new observation that quenching is generally less in an otherwise duplex mismatch than it is in single-stranded DNA presents a tool to allow one to distinguish between an RNA-DNA duplex and a reannealed homoduplex. In the case of the placement of the fluorescent base analog in the template strand opposite a correctly matched base in the nontemplate strand, low fluorescence indicates the presence of either a homo- or a heteroduplex. For a parallel experiment in which the base analog in the template strand is mismatched to an incorrect base in the nontemplate strand, low fluorescence can arise only from a (correctly paired) heteroduplex. The result that quenching in a mismatched but otherwise duplex environment is less than that in single-stranded DNA is consistent with thoughts about the structures of mismatched DNA. Combined with parallel measurements in a correctly paired construct, this approach provides a powerful tool to probe for the presence of a local heteroduplex in a transcription bubble. Using both 2-aminopurine and pyrrolo-dC placed individually at specific positions along the DNA, we now present a thorough study following the transition of the initially elongating complex. In separate experiments, probes were monitored at various positions from −2 to +13. The data are summarized in Fig.6. As summarized pictorially in Fig. 6, the initially melted bubble (as reported by probes at positions −2 and −1 on the template and nontemplate strands, respectively) remains single-stranded through translocation to position +8. As the complex translocates through position +9 to position +10, the initially melted upstream edge of the bubble reanneals. This indicates that a maximal bubble size of 10–12 nucleotides is reached on translocation to position +8. The collapse of the bubble at this point suggests a dramatic reorganization of the DNA on the protein. Retention of the upstream specificity contacts (Fig. 6,yellow) as the TATA region (−4 to −1) resumes a duplex topology would further strain the template strand connectivity. Thus we predict that promoter contacts are lost on translocation beyond position +8. Through translocation to position +10, beyond the point at which the initial bubble has collapsed, probes placed in the template strand at positions +1 and +2 remain in a heteroduplex, indicating that RNA is bound to DNA all the way back to its 5′ end (the RNA has not yet begun to dissociate). As the complex progresses to position +11, the RNA at position −1 (the 5′ end) begins to peel away. On translocation to position +12, the RNA base at position +2 relative to the start site dissociates (position +2 is no a longer heteroduplex), and the template strand base at position +1 returns to a DNA-DNA duplex. These results suggest a smooth and sequential peeling off of the RNA and indicate that the DNA bubble collapses very close behind the exiting RNA, consistent with maintaining optimal melting energetics. At this point, the heteroduplex is expected to be ∼10 base pairs in length (which, with data to date, appears to be an upper limit for the size of the heteroduplex). Interestingly, similar probes of complexes stalled near position +15 suggest a maximal heteroduplex length of 8 (9Huang J. Sousa R. J. Mol. Biol. 2000; 303: 347-358Crossref PubMed Scopus (68) Google Scholar, 11Liu C. Martin C.T. J. Mol. Biol. 2001; 308: 465-475Crossref PubMed Scopus (146) Google Scholar), suggesting a minor transition in the nature of the bubble between translocational positions +11 and +15. The introduction of a mismatched base pair at position +1 or +2 raises the possibility that the peeling away of the 5′ end of the RNA, if driven by reannealing of the DNA duplex, might occur later (at a slightly longer length of heteroduplex) in these constructs than in normal transcription. Preliminary results 3C. Liu and C. T. Martin, unpublished results. indicate a similar peeling away on constructs completely lacking the nontemplate strand, suggesting that heteroduplex length is limited by other factors. These results for the single-subunit T7 RNA polymerase are consistent with those from some recent studies of the more complex multisubunit enzymes. The elongation complexes of bothE. coli RNA polymerase and yeast Pol II are thought to contain a 8–9-bp hybrid, with the bubble melted only slightly ahead of the stalled complex (2Gnatt A.L. Cramer P. Fu J. Bushnell D.A. Kornberg R.D. Science. 2001; 292: 1876-1882Crossref PubMed Scopus (744) Google Scholar, 28Nudler E. Mustaev A. Lukhtanov E. Goldfarb A. Cell. 1997; 89: 33-41Abstract Full Text Full Text PDF PubMed Google Scholar, 29Kireeva M.L. Komissarova N. Waugh D.S. Kashlev M. J. Biol. Chem. 2000; 275: 6530-6536Abstract Full Text Full Text PDF PubMed Scopus (188) Google Scholar, 30Korzheva N. Mustaev A. Kozlov M. Malhotra A. Nikiforov V. Goldfarb A. Darst S.A. Science. 2000; 289: 619-625Crossref PubMed Scopus (338) Google Scholar), very similar to the elongation complex in T7 RNA polymerase revealed by recent fluorescence and biochemical studies (9Huang J. Sousa R. J. Mol. Biol. 2000; 303: 347-358Crossref PubMed Scopus (68) Google Scholar, 11Liu C. Martin C.T. J. Mol. Biol. 2001; 308: 465-475Crossref PubMed Scopus (146) Google Scholar).