Drosophila Factor 2, an RNA Polymerase II Transcript Release Factor, Has DNA-dependent ATPase Activity*

Drosophila factor 2 has been identified as a component of negative transcription elongation factor (N-TEF) that causes the release of RNA polymerase II transcripts in an ATP-dependent manner (Xie, Z. and Price D. H. (1996)J. Biol. Chem. 271, 11043–11046). We show here that the transcript release activity of factor 2 requires ATP or dATP and that adenosine 5′-O-(thiotriphosphate) (ATPγS), adenosine 5’-(β,γ-imino)triphosphate (AMP-PNP), or other NTPs do not support the activity. Factor 2 demonstrated a strong DNA-dependent ATPase activity that correlated with its transcript release activity. At 20 μg/ml DNA, the ATPase activity of factor 2 had an apparentK m (ATP) of 28 μm and an estimated K cat of 140 min−1. Factor 2 caused the release of nascent transcripts associated with elongation complexes generated by RNA polymerase II on a dC-tailed template. Therefore, no other protein cofactors are required for the transcript release activity of factor 2. Using the dC-tailed template assay, it was found that renaturation of the template was required for factor 2 function.

Control of transcription elongation has been recognized as an important mechanism regulating eucaryotic gene expression (1)(2)(3). The strategy to achieve the elongation control is to first apply a negative elongation potential to block and ultimately terminate RNA polymerase II transcription. The premature termination is then regulated by positive factors which enable RNA polymerase II to read through the block to transcription elongation. Based on our in vitro studies of Drosophila RNA polymerase II transcription, we proposed a model for the control of elongation by RNA polymerase II that incorporated the function of both negative and positive factors (4,5). According to the model, all RNA polymerase II molecules that initiate from a promoter are destined to produce only short transcripts because of the action of negative transcription elongation factor (N-TEF). 1 Escape from this abortive elongation into productive elongation requires the action of positive transcription elongation factor (P-TEF). P-TEFb, one of the components of P-TEF, has been purified and was found to phosphorylate the C-terminal domain of the largest subunit of RNA polymerase II (6,7). Factor 2, the first identified component of N-TEF, has been purified to apparent homogeneity as a monomer with a molec-ular mass of 154 kDa (8). Factor 2 associates with early elongation complexes and causes premature termination by releasing RNA polymerase II transcripts generated from early elongation complexes in an ATP-dependent manner (8).
ATP has been found to be required for factor-dependent termination in several transcription systems. Autoantigen La, an RNA polymerase III transcription termination factor, hydrolyzes ATP to provide energy for the termination reaction (9). The ATPase activity of La was dependent on RNA binding and was only activated by double-stranded RNA or DNA-RNA hybrid (9). In vaccinia virus, transcription termination mediated by virus-encoded termination factor (VTF) also requires ATP hydrolysis, though VTF itself is not directly responsible for the ATP hydrolysis. Instead, VTF is required to recognize a special sequence in the nascent RNA and transduce the signal to an activity named factor X, which catalyzed the ATP hydrolysis required for transcript release (10). Rho-dependent transcription termination in Escherichia coli requires ATP hydrolysis catalyzed by rho, an RNA-dependent ATPase (11)(12)(13). Rho binds to RNA and translocates along nascent transcripts to track elongating complexes and catalyzes transcript release by hydrolyzing ATP. Though ATP is required for the transcript release activity of factor 2, the role of the nucleotide and its functional relationship with factor 2 is not clear.
To understand the mechanism of factor 2 function, in this study, we investigated the importance of ATP hydrolysis in the transcript release reaction and tested for the potential ATPase activity of factor 2. In addition, we examined the transcript release activity of factor 2 using a defined system based on an immobilized dC-tailed DNA template and pure RNA polymerase II.

EXPERIMENTAL PROCEDURES
Transcript Release Assays-The transcript release assay using an immobilized Drosophila actin 5C template was as described (5,8). Transcript release assays using an immobilized 3025 dC-tailed template utilized the same protocol, except that the generation of elongation complexes was as described (14). RNase H (15) was added to dC-tailed template assays to cause template renaturation as described in the text. To examine the sensitivity of RNA transcripts to RNase A or RNase H, isolated elongation complexes were incubated with the indicated nucleases for 15 min at 25°C and then stopped with 18 l of HKE (20 mM HEPES, pH 7.6, 60 mM KCl, and 10 mM EDTA). To test the effect of ␣-amanitin on the transcript release activity of factor 2, high salt-washed early elongation complexes were first incubated with 2 g/ml ␣-amanitin at room temperature for 2 min before factor 2 and ATP or dATP were added to the reaction mixture. Labeled transcripts were quantitated using a Packard InstantImager. Assays used factor 2 (0.7 M) from glycerol gradient fraction 15 described earlier (8).
ATPase Assay-The ATPase activity of Drosophila factor 2 was assayed by measuring the release of inorganic phosphate (P i ) from [␥-32 P]ATP (Amersham Life Science, Inc.). The ATPase assay (8 l) contained 20 mM HEPES (pH 7.6), 5 mM MgCl 2 , 0.2 mg/ml bovine serum albumin, 2 mM dithiothreitol, 1 Ci of [␥-32 P]ATP, 5 M ATP, with indicated amount of factor 2 and double-stranded Act5C (4) DNA template linearized with HpaI. The reactions were initiated by the addition of 2 l of 4 ϫ label mixture (containing 80 mM HEPES, 20 mM MgCl 2 , * This work was supported by National Institutes of Health Grant RO1-GM35500. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ‡ To whom correspondence should be addressed. Tel.: 319-335-7910; Fax: 319-335-9570; E-mail: david-price@uiowa.edu. 1 The abbreviations used are: N-TEF, negative transcription elongation factor; P-TEF, positive transcription elongation factor; ATP␥S, adenosine 5Ј-O-(thiotriphosphate); AMP-PNP, adenosine 5'-(␤,␥-imino)triphosphate; VTF, virus-encoded translation factor; PEI, polyethyleneimine; nt, nucleotide(s); TRCF, transcription repair coupling factor. 8 mM dithiothreitol, 1 Ci of [␥-32 P]ATP, 20 M ATP) and incubated for 20 min at 25°C. The reactions were terminated by the addition of 1 l of 0.5 M EDTA and placed on ice. The reaction mixture (1 l) was spotted on polyethyleneimine (PEI) thin layer chromatography plate (E. Merck). The thin layer chromatography plate was developed by using 1 M formic acid and 0.5 M lithium chloride as the running buffer. The percentage of the ATP hydrolyzed was determined by quantitating the amount of the released ␥-32 P i and the unhydrolyzed [␥-32 P]ATP using a Packard InstantImager.
To determine the K m for the ATPase activity of factor 2, initial rates were measured at different concentrations of substrate ranging from 5 to 600 M ATP. The ATPase assays to measure the initial rates were carried out in the presence of 20 g/ml of HpaI linearized actin Act5C transcription template (4). The initial rates (pmol of ATP/min) versus the substrate ATP concentration (M) were plotted, and the apparent K m value was obtained from the double reciprocal plot of 1/V versus 1/[ATP].

RESULTS
Only ATP and dATP Support Transcript Release Activity of Factor 2-Previous work showed that ATP was required for Drosophila factor 2 to cause release of transcripts by RNA polymerase II (8). Here we examined the nucleotide specificity for the transcript release activity. Using an immobilized DNA fragment containing the Act5C promoter, early elongation complexes were generated and washed with 1 M KCl solution to remove the associated factor 2 ( Fig. 1, diagram at top). The high salt-washed complexes were incubated with the indicated nucleotide in the presence or absence of purified factor 2 for 5 min. Released transcripts were separated from the templateassociated transcripts by magnetic concentration. Only ATP and dATP supported the transcript release activity, while GTP, CTP, UTP, dGTP, dCTP, or dTTP did not (Fig. 1A). The nonhydrolyzable ATP analogs ATP␥S or AMPPNP did not support the transcript release activity (Fig. 1A), suggesting that the hydrolysis of ATP was required. ␣-Amanitin, which binds to RNA polymerase II and inhibits transcription elongation (16), was also examined for its effect on the activity of factor 2. The isolated elongation complexes were first incubated with ␣amanitin and then treated with factor 2 and the indicated nucleotide (Fig. 1A). ␣-Amanitin did not affect the ability of factor 2 to release RNA transcripts in the presence of either ATP or dATP (Fig. 1A). Though both ATP and dATP can be utilized by factor 2 to release RNA transcripts, 10 M ATP was utilized with the same efficiency of 300 M dATP (Fig. 1B). It is not likely that the activity seen with dATP is due to contamination with ATP since the dATP contains less than 0.5% ATP (information provided by Phamarcia Biotech).
Purified Factor 2 Has a DNA-dependent ATPase Activity-Since ATP hydrolysis was required for the function of factor 2, we determined if factor 2 had an ATPase activity. A standard thin layer chromatography assay that separated free [ 32 P]phosphate from ATP was performed, and factor 2 alone was found to have a low level of ATPase activity (Fig. 2, A and B). Previous experiments indicated that factor 2 associated with early elongation complexes (8) and had a strong affinity to DNA cellu-FIG. 1. Nucleoside triphosphate specificity for transcript release activity of factor 2. High salt-washed early elongation complexes were formed on an immobilized Act5C template as diagrammed above the gel and were then incubated with 4 nM factor 2 and the indicated nucleotides. Transcripts released from the immobilized template were analyzed as indicated on a denaturing 18% polyacrylamide gel. The first lanes of gel show the starting materials (total*) for the factor 2 incubation. A, 600 M indicated nucleotides were used in the assay. ␣-am, ␣-amanitin with a final concentration of 2 g/ml. B, titration of ATP and dATP.
FIG. 2. DNA-dependent ATPase activity of purified factor 2. A, the ATPase activity of purified factor 2 was determined by the release of 32 P i from [␥-32 P]ATP in the absence or presence of increasing amounts of double-stranded DNA (dsDNA) as indicated. The concentration of factor 2 in the reaction was around 6 nM. The reaction mixture was analyzed by PEI cellulose thin layer chromatography and detected by using a Packard InstantImager. B, graphic representation of the percentage of ATP hydrolysis of reactions in panel A. C, titration of factor 2 (0, 1 , 2, 6, 20, 60, and 120 nM) in the ATPase assay at a fixed concentration of 5 g/ml DNA. D, graphic representation of the percentage of ATP hydrolysis in panel C.
lose, 2 suggesting that factor 2 might bind DNA. To see if DNA might activate the ATPase activity of factor 2, reactions were supplemented with increasing amounts of DNA. At a fixed concentration of 6 nM factor 2, ATPase activity was dramatically stimulated by DNA and did not plateau even when the concentration of DNA was increased up to 80 g/ml (Fig. 2, A  and B). The DNA-dependent ATPase activity of factor 2 required a divalent cation with Mg 2ϩ or Mn 2ϩ functioning equally well (data not shown). When the DNA concentration was fixed at 5 g/ml, increasing factor 2 caused a linear increase in ATPase activity until the percent hydrolysis was so great that ATP became limiting during the reaction (Fig. 2, C and D). Although our previous results suggested that factor 2 may associate with DNA (8), the lack of saturation with either DNA or factor 2 suggests that there is no stable interaction with DNA in the presence of ATP. Since dATP also supported transcript release, we examined the ability of factor 2 to hydrolyze [␣-32 P]dATP to dADP. dATPase activity was detected and, consistent with the reduced transcript release activity supported by dATP, was about 5% of the ATPase activity.
To provide further evidence that the DNA-dependent ATPase was associated with factor 2, an ATPase assay was performed across the glycerol gradient fractions of the final step of factor 2 purification (8). Results from both an ATPase assay and a transcript release assay across the same set of fractions (data from Ref. 8) were quantitated and compared (Fig. 3). The ATPase correlated with the transcript release, suggesting that the DNA-dependent ATPase activity was in-trinsic to factor 2.
Primary kinetic parameters for the ATPase activity of factor 2 were also determined under conditions similar to that used for transcription except that only ATP was used. The apparent K m(ATP) for factor 2 was determined by measuring the initial velocity of ATP hydrolysis at different concentrations of ATP (Table I). A double reciprocal plot (1/V versus 1/[ATP]) of the data (Fig. 4) yielded an apparent K m(ATP) for factor 2 of 28 M and a V max of 6.25 pmol of ATP hydrolyzed/min. After estimating the amount of factor 2 based on the absorbance at 280 nM (A 280 ), the turnover number K cat was calculated to be about 140 min Ϫ1 . These results indicated that factor 2 is very active in hydrolyzing ATP under conditions normally used for transcription.
Factor 2 Causes Transcript Release from Elongation Complexes Formed on a dC-tailed Template-To determine if factor 2 was competent to function as a transcript release factor without additional protein cofactors, we used a dC-tailed template that allows efficient transcription initiation by pure RNA polymerase II. As diagrammed in (Fig. 5, top), initiation of prebound RNA polymerase II was accomplished with a 30-s pulse. The labeled complexes were isolated and then chased to generate transcripts between 50 and 250 nt in length. The resulting elongation complexes containing transcripts longer than 20 nt were examined for their sensitivity to factor 2 (transcripts less than 20 nt were analyzed separately and will be discussed below). When the isolated elongation complexes were incubated with only buffer or ATP, less than 10% of the transcripts were released (Fig. 5A). When the reactions were supplemented with factor 2 and ATP, an additional 17% of the nascent transcripts were released (Fig. 5A). The transcripts resistant to release by factor 2 did not result from the use of a subsaturating level of factor 2 since increasing the amount of factor 2 by 3-fold did not cause further release (Fig. 5A). These results suggested that factor 2 could function alone with the caveat that most of the transcripts were resistant to factor 2.
Although it was possible that a protein cofactor might be required to achieve complete release of transcripts in the dCtailed template assay, it was more likely that the factor 2resistant complexes resulted from an unusual feature of transcription of dC-tailed templates. It has been found that most of the transcripts generated by Drosophila RNA polymerase II are in heteroduplex with the template strand (RNase H sensitive), which prevents the renaturation of the template (15). As expected, RNase H degraded 80% of the transcripts associated with elongation complexes (beads) but did not affect the small amount of transcripts found in the supernatant fraction (Fig.  5B). Consistent with this result, less than 20% of the transcripts associated with elongation complexes were sensitive to RNase A, whereas the transcripts in the supernatant were completely sensitive (Fig. 5B). The fraction of transcripts released from elongation complexes by factor 2 was similar to the fraction of transcripts sensitive to RNase A or resistant to 2 Z. Xie and D. Price, unpublished results.  RNase H. This suggested that factor 2 caused release of only transcripts that were not in heteroduplex. Indeed, the transcripts found in factor 2-resistant complexes were sensitive to RNase H (last two lanes in Fig. 5B). The transcripts around 13 nt in length were resistant to factor 2 function (Fig. 5A) and were also resistant to RNase H and RNase A digestion (Fig. 5B). In a complementary experiment, elongation complexes generated in the presence of RNase H were examined for sensitivity to factor 2. To confirm that the transcripts were no longer in heteroduplex, the complexes were incubated with either RNase H or RNase A. The transcripts present in these elongation complexes were not sensitive to RNase H but were degraded by RNase A (Fig. 6A). When these elongation complexes were incubated with only ATP, a low level (about 15%) of transcripts was released in the supernatant fraction, which is similar to the level of transcript release during the RNase H incubation (Fig. 6B). When factor 2 was added along with ATP, transcript release increased to 35% (Fig. 6B). When the level of factor 2 was tripled, transcript release was increased to a total of more than 80% (Fig. 6B). Transcript release was dependent on ATP as found before (data not shown). These results indicate that no other protein factors were required for the transcript release activity of factor 2 and that only transcripts displaced from the template strand were released by factor 2.

DISCUSSION
Factor 2 has been shown to promote the dissociation of RNA polymerase II transcripts from ternary elongation complexes in an ATP-dependent manner (8). To understand how factor 2 causes transcript release and ultimately elucidate its role in elongation control, we investigated the biochemical properties of the factor in defined in vitro systems. With an immobilized template assay, we found that hydrolysis of ATP was required for the transcript release activity of factor 2 and that only dATP could substitute for ATP. Using pure factor 2, we found that it exhibited an ATPase activity that was strongly dependent on DNA. We employed a dC-tailed template assay with pure RNA polymerase II and found that factor 2 did not require any other protein cofactors. In addition, we demonstrated that transcript release occurred only under conditions that caused transcript displacement and template renaturation.
Even though factor 2 shares a requirement for ATP hydrolysis with other termination factors, our results suggest that factor 2 utilizes a different mechanism for transcript release. ATP hydrolysis is required for termination by vaccinia VTF (17,18), E. coli rho (11)(12)(13), and RNA polymerase III termina- tion factor La (9,19). However, the ATPase activity of factor 2 is activated by double-stranded DNA, whereas that of rho and La are only stimulated by RNA or RNA containing nucleic acids but not double-stranded DNA. Furthermore, unlike the other factors, association of factor 2 with a specific sequence in the nascent RNA is not required for its transcript release activity. Factor 2 associated with early elongation complexes containing RNA less than 10 nucleotides in length and caused transcript release (8). Such short transcripts are probably still sequestered within RNA polymerase II (20,21), and thus factor 2 is unlikely to translocate along the nascent transcript to locate paused elongation complexes. Instead, factor 2 is more likely to find elongation complexes through an interaction with the template and hydrolyze ATP to provide energy to disrupt the interaction of nascent RNA with RNA polymerase II.
Though DNA strongly activates the ATPase activity of factor 2, the mechanism of the activation is not clear. Nucleic acids have been shown to play an important role in activating various ATPases. RNA is an essential cofactor for the ATPase rho. RNA stabilizes rho as a hexamer (22) and lowers its sensitivity to trypsin digestion (23), suggesting that RNA may activate its ATPase activity by inducing conformational change of rho. DNA has also been shown, for example, to activate the ATPase activity of E. coli Rep protein, by inducing dimerization (24,25), or Rec A, by causing conformational changes (26). Factor 2 is a monomer of 154 kDa with little ATPase activity in the absence of DNA. It is possible that DNA induces a conformational change or causes oligomerization of factor 2 and, thereby, stimulates its ATPase activity.
The characterization of factor 2 function using a dC-tailed template assay revealed that transcript release was blocked when the RNA remained in heteroduplex with the template. It is possible that extensive heteroduplex formation stabilizes the association of the RNA with the template and thus increases the energy required to cause transcript release beyond that provided by factor 2. However, another more interesting explanation is that double-stranded DNA, not RNA/DNA hybrid, is required upstream of elongation complexes for factor 2-mediated transcript release. This possibility is suggested by the finding that the 13-mer generated on a dC-tailed template was resistant to the transcript release activity of factor 2 (Fig. 5A). Factor 2 has been shown to dissociate transcripts as short as 9 or 10 nucleotides in length generated from promoter containing templates (8). It is possible that the resistance of 13-mer to factor 2 is due to the lack of double-stranded DNA upstream of the ternary complex. However, previous studies suggested that the polymerases with nascent 13-mers had properties of the arrested state (27,28), which raised a possibility that the insensitivity of 13-mers to factor 2 action might be caused by the arrested state of the polymerase. To examine this possibility on more normal arrested complexes, an immobilized template was used to generate complexes arrested 112 base pairs downstream of the Act5C promoter (4,8). These complexes are not able to elongate in the presence of 600 M nucleotide triphosphates but are sensitive to DmS-II action (4). When these complexes were incubated with factor 2 in the presence of ATP, more than 60% of the transcripts were released (data not shown). The elongation complexes blocked at the same site were also observed to be sensitive to factor 2 action during elongation (Fig. 4B in Ref. 8). Thus, factor 2 can release transcripts associated with arrested elongation complexes. The insensitivity of 13-mer to factor 2 is likely caused by the physical state of DNA upstream of the elongation complex. Doublestranded DNA may be required for appropriate localization of the ATPase activity of factor 2 and may enable its coupling with transcript release. Since the elongation complexes formed on dC-tailed template were generated by pure RNA polymerase II, a particular localization of factor 2 in relationship to the elongation complex may be essential for a specific interaction between factor 2 and RNA polymerase.
How factor 2 activity is directed toward a subset of the total complexes is not understood. The transcript release activity of factor 2 per se does not need other protein cofactors. However, factor 2 may be modulated by other factors in the process of transcription elongation. An inhibitory activity of factor 2 was found to associate with low salt-washed early elongation complexes and protect a fraction of the complexes from being released by factor 2 (8). Furthermore, in reactions containing the factors required for the generation of DRB-sensitive long runoff transcripts, the appearance of DRB-sensitive runoff transcripts were suppressed by increasing the amount of purified factor 2 2 but not by a partially purified factor 2 fraction (6). This finding suggests that factor 2 has the ability to terminate the potentially productive elongation complexes that are going to generate DRB-sensitive runoff transcripts. The antitermination activity of factor 2 associated with the low salt-washed early elongation complexes or contained in the partially purified factor 2 fraction is likely required to restrict factor 2 from acting on the potentially productive elongation complexes. A preliminary result suggests that a 14-kDa protein that copurified with factor 2 through multiple steps, but was removed during chromatography on hydroxyapatite column, is required to support the production of DRB-sensitive runoff transcripts in the presence of purified factor 2. 2 Thus, during transcription elongation, the transcript release activity of factor 2 is likely to be regulated by other factors so that factor 2 preferentially acts on the elongation complexes that are not rescued by P-TEF.
In addition to a role in transcription elongation, factor 2 exhibits several of the properties of the E. coli transcription repair coupling factor (TRCF) (29). Like TRCF, factor 2 has an associated ATPase activity (this paper) and causes the transcript release from stalled elongation complexes (8). Recently, ERCC6, a putative human homologue of TRCF, has been found FIG. 6. Transcript release activity of factor 2 on a renaturing dC-tailed template. A, RNase A and RNase H sensitivity of transcripts. Complexes were generated by pulse labeling, chasing in the presence of RNase H, and low salt isolation and were then incubated with RNase A or RNase H for 15 min. B, transcript release assay. Complexes were generated as in panel A and subjected to the transcript release assay. S, released transcripts; B, bound transcripts; total*, total transcripts. All transcripts were analyzed by denaturing 18% polyacrylamide gel electrophoresis.
to have an ATPase activity that can be activated by DNA. However, ERCC6 lacks the ability to terminate stalled RNA polymerase II (30). It was suggested that eucaryotes might utilize a different mechanism in transcription-coupled repair than E. coli in which the polymerase backs away from the damaged DNA due to the action of S-II (30,31). It is not clear why such a difference should exist since similar transcript cleavage factors GreA and GreB exist in E. coli (32). In support of this idea, it has recently been shown that S-II is not required for transcription-coupled repair in yeast (33). We suggest the possibility that factor 2 may function as a TRCF, but further examination of the role of factor 2 in repair will be necessary to clarify its cellular function.