INTRODUCTION

A role for an ATP cofactor in eukaryotic messenger RNA synthesis was first brought to light by Weinmann and co-workers (1), who observed that AMP-PNP¹ could not replace ATP in synthesis of promoter-specific transcripts by RNA polymerase II from the AdML promoter, even though AMP-PNP is a substrate for elongation by polymerase. In subsequent experiments, Luse and Jacob (2) demonstrated that ATP is required for synthesis of the first phosphodiester bond of nascent transcripts initiated by RNA polymerase II at the AdML promoter. These findings have been confirmed and extended in work from several laboratories (3, 4, 5, 6). We identified ATPS as a potent inhibitor of the ATP-requiring step in initiation by RNA polymerase II, and we have used ATPS to obtain evidence that ATP is utilized prior to initiation to promote conversion of the fully assembled, but inactive, preinitiation complex into a transcriptionally active conformation in a reversible step that results in formation of a transient activated preinitiation complex and decays to an inactive state in the presence of ATPS with a t1/2 of ~40 s (5, 7). Gralla and co-workers have shown that ATP is utilized at least in part for formation of an ``open'' promoter complex (8, 9), and Timmers and co-workers (6, 10, 11) and others (12, 13, 14) have obtained strong circumstantial evidence that ATP-dependent open complex formation is catalyzed by the TFIIH DNA helicase. Notably, ATP-dependent open complex formation, like the ATP-dependent activation step in transcription is susceptible to inhibition by ATPS (6).

Whether an ATP cofactor is also required after initiation is unclear. Sawadogo and Roeder (15) showed that ATP is not needed after the synthesis of 9-10-nt transcripts initiated from the AdML promoter, and Luse and co-workers (16, 17) reported that short, 4-nt transcripts initiated from the same promoter can be extended to 10-12-nt products in the absence of ATP. Based on indirect evidence, it was recently proposed that ATP hydrolysis is required only after initiation to promote escape of RNA polymerase II from the promoter in an ATP-dependent step catalyzed by the TFIIH DNA helicase (18). This proposal has been difficult to reconcile, however, with the large amount of evidence indicating that ATP is required for initiation.

To gain insight into the role of ATP in early steps after transcription initiation, we have now directly investigated the ability of stably initiated RNA polymerase II transcription complexes containing short, ~5-8-nucleotide transcripts to escape the promoter in the absence of an ATP cofactor. We discovered that a significant fraction of early RNA polymerase II transcription complexes becomes arrested as a result of ATP deprivation and is unable to escape the promoter. Furthermore, we observe (i) that addition of ATP to transcription reactions prior to arrest of polymerase at these sites is sufficient to suppress arrest and (ii) that a fraction of arrested elongation complexes can be re-activated by addition of ATP. Taken together, our findings reveal a novel role for an ATP cofactor in transcription by RNA polymerase II.

EXPERIMENTAL PROCEDURES

Unlabeled ultrapure ribonucleoside 5-triphosphates and dATP were purchased from Pharmacia Biotech Inc. Dinucleotides CpA, CpU, and UpC, -amanitin, immobilized hexokinase, and polyvinyl alcohol (type II) were purchased from Sigma. ATPS was from Boehringer Mannheim. [-³²P]CTP (>400 Ci/mmol) was purchased from Amersham Corp. Bovine serum albumin (Pentex fraction V) was obtained from ICN Immunobiologicals. Recombinant ribonuclease inhibitor (RNasin I) was from Promega.

RNA polymerase II (19) and TFIIH (rat , TSK DEAE 5-PW fraction, (20)) were purified as described from rat liver nuclear extracts. Recombinant yeast TBP (AcA 44 fraction (21, 22)) and TFIIB (rat  (23)) were expressed in Escherichia coli and purified as described. Recombinant TFIIE was prepared as described (24), except that the 56-kDa subunit was expressed in E. coli strain BL21(DE3)-pLysS. Recombinant TFIIF was purified as described (25) from E. coli strain JM109(DE3) co-infected with M13mpET-RAP30 and M13mpET-RAP74.

Except as indicated in the figure legends, preinitiation complexes were assembled at the AdML promoter at 28 °C by a 45-min preincubation of 35-µl reaction mixtures containing 20 m Hepes-NaOH (pH 7.9), 20 m Tris-HCl (pH 7.9), 60 m KCl, 4 m MgCl₂, 0.1 m EDTA, 1 m dithiothreitol, 0.5 mg/ml bovine serum albumin, 2% (w/v) polyvinyl alcohol, 7% (v/v) glycerol, 6 units of RNasin, ~20 ng of the EcoRI to NdeI fragment from pDN-AdML (7), ~50 ng of recombinant yeast TBP, ~10 ng of recombinant TFIIB, ~20 ng of recombinant TFIIF, ~20 ng of recombinant TFIIE, ~150 ng of TFIIH, and ~0.01 unit of RNA polymerase II. Unless indicated otherwise in the figure legend, transcription was initiated by addition of a labeling mix containing 200 µ dinucleotide primer, 5 µ of ATP or dATP, 0.01 µ UTP, and 0.5 µ [-³²P]CTP for 10-20 min at 28 °C. Short transcripts synthesized during the labeling phase of the reaction were chased into longer products by addition of 200 µ of unlabeled CTP and other ribonucleoside triphosphates as indicated in the figure legends. For resolution of short transcripts, 15 µl of the reaction mixture was added to 6 µl of a stop solution containing 100 m EDTA and 0.5 mg/ml proteinase K. Following incubation at room temperature for 15 min, 25 µl of a urea-dye solution containing 10  urea, 0.025% bromphenol blue, and 0.025% xylene cyanol FF were added. The samples were vortexed for 10 s, heated at 70 °C for 5 min, and separated on a 25% acrylamide, 3% bis, 7.0  urea gel as described (26). For resolution of 254-nt run-off transcripts, an equal volume of a stop solution containing 25 m EDTA, 2% SDS, 200 m Tris-Cl (pH 7.6), 300 m NaCl, and 0.5 mg/ml proteinase K was added to the reaction mixture. Following a 15 min incubation at room temprature, RNA was precipitated by adding 2.5 volumes ethanol, incubating 10 min at 70 °C and spinning in a microcentrifuge for 20 min at 12,000 × g. The RNA pellet was washed with 70% ethanol, dried briefly in a Speed-Vac, and resuspended in 25 µl of the urea-dye solution. Samples were separated on 6% acrylamide, 0.8% bis, 7.0  urea gels. Gels were imaged by autoradiography or on a Molecular Dynamics PhosphorImager.

2.5 ml AcA 34 gel filtration resin was packed into a Pasteur pipette fitted with a glass-wool plug and equilibrated with a transcription buffer solution identical to the transcription reaction mix, excluding the DNA, transcription factors, and nucleotides. Transcription reactions were performed at five times the regular size. Following synthesis of short transcripts in the labeling phase of the reaction, reaction mixtures were loaded onto the column at room temprature and eluted with transcription buffer. Fractions of 100 µl were collected.

Hexokinase-agarose beads were equilibrated with transcription buffer and resuspended to make a 1:1 slurry. After the labeling phase of transcription was completed, 15 µl of the slurry (~2.5 units of hexokinase) and 2.0 µl of 100 m dextrose were added to each 30-µl reaction for an additional 15 min. incubation at 28 °C. The hexokinase beads were then removed by centrifugation for 30 s in a microcentrifuge, and the supernatant was used in the chase phase of transcription reactions. A similar procedure was used to deplete ATP from chase nucleotides.

RESULTS AND DISCUSSION

We observed previously that ATPS is a potent inhibitor of the ATP-requiring step in TFIIH-dependent transcription initiation by RNA polymerase II (5, 7). To investigate the possibility that an ATP cofactor is required at a post-initiation stage of transcription, we tested the effect of ATPS on extension of 5-8-nt transcripts contained in promoter-proximally paused RNA polymerase II elongation complexes initiated from the AdML promoter.

In the experiment of Fig. 1B, promoter-proximally paused elongation complexes were formed in a basal transcription system composed of RNA polymerase II, TFIIH, and recombinant TBP, TFIIB, TFIIE, and TFIIF. Short radioactive transcripts were synthesized by incubating preassembled preinitiation complexes for 10 min with 200 µ of the initiating dinucleotide CpU, 5 µ ATP, 10 n UTP, and 0.5 µ [-³²P]CTP. Under these conditions, transcripts with a maximum length of 5-8 nt were synthesized (lane 1); synthesis of these transcripts was inhibited by 1 µg/ml -amanitin (data not shown). A significant fraction of these products were tri- or tetranucleotide abortive transcripts, which could not be chased into longer products. The stably initiated transcripts were then chased into longer products in the presence of ATP (lanes 2-6) or ATPS (lanes 7-11) by addition of 200 µ CTP, 100 µ UTP, and 100 µ of the RNA chain-terminating nucleotide 3-O-MeGTP, which prevents most transcription beyond the first G in the AdML transcript. Because ATP is not incorporated into CpU-initiated transcripts between positions 4 and 17, any differences in the efficiency of elongation in the presence and absence of ATP must reflect a role for ATP other than as a substrate for RNA synthesis by RNA polymerase II.

ATPS induces arrest during the extension of promoter-proximally paused transcription complexes.

AS

AS

In the presence of ATP, most of the stably initiated transcripts were rapidly chased into 18-nt 3-O-MeG-terminated RNAs. In the presence of ATPS, most of the short transcripts were rapidly chased into 9-13-nt products, and synthesis of the full-length 3-O-MeG-terminated transcript was dramatically reduced, reaching a maximum level within 10 min of addition of the chase nucleotides.

The ability of ATPS to inhibit elongation by promoter-proximally paused transcription complexes did not depend on the particular initiating dinucleotide, since ATPS also inhibited extension of short transcripts initiated with ATP and with the dinucleotides UpC and CpA (Fig. 1C). It is noteworthy that the distribution of products formed in the presence of ATPS varies depending on the initiating dinucleotide, suggesting that these products do not accumulate a fixed distance from the site of initiation. In addition, we observed that ATPS inhibited elongation by promoter-proximally paused transcription complexes when the labeling phase of the reaction was limited to 1 min, thereby reducing the length of the pause (data not shown).

Consistent with previous results indicating that ATP is not required for elongation by RNA polymerase II under most conditions (1, 3, 7, 15, 27), the inhibitory effect of ATPS is observed only on very early elongation complexes. In the experiment of Fig. 2, the concentration of nucleotides present in the labeling phase of the reaction were varied in order to allow synthesis of different length transcripts prior to addition of chase nucleotides and ATP or ATPS. Whereas extension of 5-8-nt transcripts into 18-nt, 3-O-MeG-terminated (lanes 1-6) or 254-nt run-off transcripts (lanes 10-12) was significantly reduced in the presence of ATPS, 9-10 nt and longer transcripts were chased into 254-nt run-off products with similar efficiency in the presence of either ATP or ATPS (lanes 7-9, 13-15).

Transcription complexes that have synthesized more than 9-10-nt RNAs are not sensitive to inhibition by ATPS.

AS

The 9-13-nt products synthesized in the presence of ATPS could be either terminated and released or associated with arrested, but potentially active, elongation complexes. To address this question, we asked whether 9-13-nt transcripts synthesized in the presence of ATPS could be chased into longer products following addition of ATP. As shown in Fig. 1B, a significant fraction of ATPS-inhibited RNA polymerase II elongation complexes could be chased into 3-O-MeG-terminated RNA products when ATP was added at a 5-fold molar excess over ATPS (lanes 12-14), indicating that they were arrested, but potentially active,elongation complexes.

Taken together, these results indicate that ATPS can induce arrest by very early RNA polymerase II elongation complexes, and they suggest a role for an ATP cofactor in suppression of arrest by RNA polymerase II at promoter-proximal sites. To test this possibility directly, promoter-proximally paused elongation complexes containing short CpU-initiated transcripts were purified by gel filtration to remove ATP and unincorporated nucleotides. Transcripts associated with purified elongation complexes were then chased into longer products by addition of UTP, CTP, and 3-O-MeGTP in either the presence or absence of ATP. As shown in Fig. 3A, in the presence of ATP, nearly all of the short transcripts were chased into 3-O-MeG-terminated products. In the absence of added ATP, however, only a small fraction of the short transcripts could be extended into 16 nt, U-terminated products; the remaining transcripts were paused or terminated following synthesis of 10-14-nt products.

As an alternative method to remove ATP from reaction mixtures, we treated promoter-proximally paused elongation complexes with immobilized hexokinase, which hydrolyzes ATP in the presence of glucose to give ADP and glucose 6-phosphate. Consistent with results from the gel filtration experiment, nearly all of the short transcripts were chased into 3-O-MeG-terminated products in the presence of ATP, whereas, in the absence of added ATP, only a small fraction of short transcripts were successfully extended into the 16-nt, U-terminated product (Fig. 3B). We note that a very small fraction of transcripts were elongated past the U residue at position 16, suggesting that we have been unable to remove all ATP from reaction mixtures by gel filtration and hexokinase treatment. We do not know whether all of the transcription complexes would have become arrested during extension of 5-8-nt transcripts if we were able to remove ATP from reaction mixtures entirely.

It is well established that TFIIH-dependent transcription by RNA polymerase II requires a hydrolyzable ATP cofactor for synthesis of the first phosphodiester bond of nascent transcripts. Our findings indicating that very early RNA polymerase II elongation complexes are susceptible to arrest in either the absence of ATP or in the presence of ATPS bring to light a new role for an ATP cofactor in promoter escape by RNA polymerase II.

ATP-dependent activation of the preinitiation complex results in formation of a transiently activated intermediate that decays relatively slowly to an inactive state with a t1/2 of ~40 s (5, 7). Under most conditions, this interval is sufficient for synthesis of stably initiated RNA polymerase II elongation complexes containing transcripts greater than 20 nucleotides in length (e.g. Refs. 28 and 29).² In the experiments presented here, it was necessary to separate the ATP-dependent activation step in initiation from subsequent ATP-dependent events. We therefore studied RNA polymerase II elongation complexes that were initiated in the presence of ATP but forced to pause at promoter-proximal sites prior to depletion of ATP. As a consequence, our experiments did not address the question whether a single ATP activation event prior to synthesis of the first phosphodiester bond can be sufficient for both initiation and escape of RNA polymerase II from the promoter.

In an effort to address this question, we carried out an ATPS challenge experiment. In this experiment, RNA polymerase II preinitiation complexes were assembled at the AdML promoter, preactivated by incubation with dATP for 1 min, and then treated with excess ATPS to block further ATP activation events. At various times after addition of ATPS, transcription was initiated by addition of CpA, CTP, UTP, and GTP, which are sufficient for synthesis of full-length run-off transcripts. As shown in Fig. 4, full-length run-off transcripts were synthesized under these conditions, indicating that a single ATP activation event prior to synthesis of the first phosphodiester bond can be sufficient for both initiation and escape of RNA polymerase II from the promoter. It is important to note, however, that we do not know what fraction of initiated transcription complexes were able to escape arrest, since, in the presence of ATPS, RNA polymerase II can initiate transcription only prior to decay of the activated preinitiation complex. Under these conditions, the level of initiation is too low to allow reliable measurement of short transcripts. Thus, although these experiments indicate that at least some fraction of transcription complexes were able to initiate and escape the promoter without a second ATP-dependent activation event, they do not rule out the possibility that a second ATP activation event contributes to the efficiency with which unpaused transcription complexes escape the promoter.

In summary, in this report we have investigated the role of ATP in post-initiation stages of transcription by RNA polymerase II in a basal transcription system reconstituted with RNA polymerase II, TBP, TFIIB, TFIIE, TFIIF, and TFIIH by investigating the ability of stably initiated RNA polymerase II elongation complexes containing short, ~5-8-nucleotide transcripts to escape the promoter in the absence of an ATP cofactor. Although our findings argue that an ATP cofactor is not essential for post-initiation stages of TFIIH-dependent transcription, our experiments led to the discovery that, in the absence of a hydrolyzable ATP cofactor, a significant fraction of early RNA polymerase II elongation complexes suffer arrest at promoter-proximal sites. Furthermore, we observe (i) that addition of ATP to transcription reactions prior to arrest of polymerase at these sites is sufficient to suppress arrest and (ii) that a fraction of arrested elongation complexes can re-activated by addition of ATP. Although it is presently not clear why RNA polymerase II elongation complexes that are forced to pause at promoter-proximal sites suffer arrest following ATP-depletion or addition of ATPS, one possible explanation is that arrest results when elongation complexes fail to escape the promoter before decay of the activated state or collapse of the open complex. Under these conditions, an additional ATP-dependent activation event, perhaps involving the TFIIH DNA helicase, may be necessary for synthesis of stably initiated, active elongation complexes containing transcripts greater than 14 nucleotides in length. Regardless of the precise mechanism, our findings reveal a novel role for an ATP cofactor in suppression of arrest by RNA polymerase II at promoter-proximal sites.