Phosphorylation of RNA Polymerase IIA Occurs Subsequent to Interaction with the Promoter and before the Initiation of Transcription*

The largest subunit of mammalian RNA polymerase II contains at its C terminus an unusual domain con- sisting of multiple tandem repeats of the seven-amino acid consensus sequence Tyr-Ser-Pro-Thr-Ser-Pro- Ser. This domain is unphosphorylated in RNA polymerase IIA and extensively phosphorylated in RNA po- lymerase 110. To investigate the role of the C-terminal domain and the functional significance of its phos- phorylation, changes in the level of phosphorylation were followed as a function of the position of RNA polymerase II in the transcription cycle. Complexes were formed with 32P-labeled RNA polymerase IIA and separated from the free polymerase by gel filtration. The phosphorylation state of the RNA polymerase II largest subunit was determined by sodium dodecyl sul- fate-polyacrylamide gel electrophoresis. Results indi- cate that RNA polymerase IIA interacts with the tem- plate-committed complex to form a stable preinitiation complex. RNA polymerase IIA associated with such complexes is converted to RNA polymerase 110 in the presence of ATP prior to the formation of the first phosphodiester bond. that purified

to the association of enzyme with the promoter and prior to the initiation of transcription.
The regulation of gene expression is determined to a major extent by the frequency with which RNA polymerase initiates the transcription of specific genes. An understanding of the mechanisms involved in gene regulation is, therefore, dependent on our understanding of the basic reactions involved in initiation.
Although there has been considerable progress in our understanding of general factors required for the expres- sion of class II genes (Van Dyke et al., 1988;Buratowski et al., 1989;Saltzman and Weinmann, 1989) and factors that influence the rate of initiation by RNA polymerase II (Ptashne, 1988;Mitchell and Tjian, 1989), relatively little is known about how these factors interact with RNA polymerase II to influence the rate of transcription.
The largest subunit of mammalian RNA polymerase II contains at its C terminus an unusual domain consisting of multiple repeats of the seven-amino acid consensus sequence Tyr-Ser-Pro-Thr-Ser-Pro-Ser (Corden et al., 1985). While this domain does not appear to be essential for the in vitro transcription of the AdB-MLP,' at least half of the repeats are required for in vivo function Zehring et al., 1988;Payne et al., 1989;Nonet et al., 1987;Allison et al., 1988;Bartolomei et al., 1988). Mammalian cells contain two forms of RNA polymerase II, designated 110 and IIA, that differ with respect to the level of phosphorylation within this domain (Kim and Dahmus, 1986;Cadena and Dahmus, 1987). The C-terminal domain of RNA polymerase subunit IIa is unmodified, whereas the C-terminal domain of subunit 110 is extensively phosphorylated (Cadena and Dahmus, 1987).
Although the function of the C-terminal domain of subunit IIa is not known, recent results indicate that the level of phosphorylation of this domain changes during the course of transcription (Bartholomew et al., 1986;Payne et al., 1989). Based on these results, we have proposed that RNA polymerase IIA, containing an unphosphorylated C-terminal domain, interacts with DNA to form a preinitiation complex and that phosphorylation occurs during the transition from initiation to elongation Payne et al., 1989). A similar model has been proposed by Sigler (1988). The crosslinking of nascent transcripts to RNA polymerase subunit 110 both in in vitro transcription systems (Bartholomew et al., 1986;Payne et al., 1989) and in isolated HeLa nuclei (Cadena and Dahmus, 1987) ATP as indicated in the figure legends. and 30 UM AMP-PNP,'UTP, and GTP. Complete transcription reactions were incubated at 30 "C for 45 min. RNA was purified and analyzed on 5% polyacrylamide-urea gels as described by Dahmus and Kedinger (1983). The amount of specific transcript was quantitated as described previously .

Fractionation of Preinitiation
Complexes and Free RNA Polymerase II by Gel Filtration-An identification of the precise step in the transcription reaction in which phosphorylation of the C-terminal domain occurs is an essential prerequisite to defining the involvement of this domain in transcription. The overall strategy is to initiate transcription with purified RNA polymerase IIA in the presence of a fractionated transcription extract and to determine the level of phosphorylation of subunit IIa as RNA polymerase progresses through the various steps of transcription. In order to assess the level of phosphorylation of RNA polymerase II in transcription complexes, it is essential to first separate these complexes from free enzyme.
Purified RNA polymerase IIA was incubated in the presence of a reconstituted HeLa cell transcription extract and a DNA template containing the AdB-MLP to form preinitiation complexes. Complexes were then separated from free RNA polymerase II by gel filtration as described under "Experimental Procedures." RNA polymerase II associated with template is excluded and elutes in the void volume, whereas free RNA polymerase is included in the column. To determine whether or not complexes contained in the excluded fractions are indeed functionally active preinitiation complexes, the following experiment was done. Preinitiation complexes were formed in the absence of ATP by incubation of RNA polymerase IIA and a reconstituted transcription extract in the presence of DNA template. The DNA template was truncated to produce a 560-nucleotide runoff transcript. Preinitiation complexes were purified by gel filtration and incubated in the presence of increasing concentrations of a competing second template, (405 template) truncated to give a runoff transcript of 405 nucleotides.
Nucleotides were then added, and, after an additional 5-min incubation, Sarkosyl was added to limit transcription to a single round (Hawley and Roeder, 1987). The complete reaction was incubated for an additional 25 min, and the transcripts formed were analyzed as described under "Experimental Procedures." Results presented in Fig.  1 (lanes 1-3) show that purified preinitiation complexes formed with the 560 template produce only a 560-nucleotide transcript even in the presence of excess 405 template. Similarly, transcription from preinitiation complexes formed on the 405 template is restricted to that template even in the presence of competing 560 template (Fig. 1, lanes 4-6). These results demonstrate that the preinitiation complexes formed with purified RNA polymerase IIA in the absence of ATP and fractionated by gel filtration are stable and able to initiate and form completed transcripts upon the addition of nucleotides. Furthermore, the fact that end to end transcription of the competing template is not observed indicates that excluded fractions do not contain significant amounts of free RNA polymerase II.

Analysis of the Phosphorylation State of RNA Polymerase II in Transcription
Complexes-The decreased electrophoretic mobility in SDS-PAGE of RNA polymerase subunit 110, relative to subunit IIa, results from extensive phosphorylation within the C-terminal domain (Cadena and Dahmus, 1987;Cisek and Corden, 1989). Changes in the phosphorylation state of the largest subunit of RNA polymerase II can therefore be monitored by following the shift in electrophoretic mobility. In order to increase the sensitivity of such assays and to facilitate the analysis of factors that catalyze the phosphorylation of RNA polymerase II, Payne et al. (1989) developed an assay based on the utilization of "'P-labeled RNA polymerase IIA as substrate. Casein kinase II phosphorylates subunit IIa at a single site that does not result in a mobility shift in SDS-PAGE (Dahmus, 1981;Payne et al., 1989). "'P-Labeled RNA polymerase IIA, labeled by phosphorylation with casein kinase II as described under "Exper- imental Procedures," was incubated in the presence of a reconstituted transcription extract, with or without template DNA and ATP, and fractionated from transcription factors and free RNA polymerase II by gel filtration as described above. The distribution of RNA polymerase II in the column fractions and the state of phosphorylation of subunit IIa was determined by SDS-PAGE. In the absence of DNA, RNA polymerase IIA is found exclusively in the included fractions (Fig. 24, lanes 5-8). The presence of DNA results in a fraction of RNA polymerase IIA eluting in the excluded fractions (Fig.  2B, lanes Z-3). The conditions for preinitiation complex formation and fractionation are identical with those employed in Fig. 1 in which excluded fractions were shown to contain template committed transcription complexes. Consequently, these results suggest that RNA polymerase IIA forms a stable preinitiation complex on the AdB-MLP in the absence of ATP.
The fractionation of free and template-bound RNA polymerase II becomes more apparent in reactions that contain ATP. The inclusion of ATP, in the absence of DNA, results in the partial conversion of RNA polymerase IIA to 110 as indicated by the reduced electrophoretic mobility of subunit IIa in included fractions (Fig. ZC, lanes 6-8). This is in agreement with previous results that demonstrate that transcription extracts contain a protein kinase active in the phosphorylation of the C-terminal domain of subunit IIa . In the presence of DNA and ATP, RNA polymerase 110 is recovered in the excluded fractions (Fig. 20, lanes  2-4). These results, in conjunction with the results presented in Fig. 1, suggest that RNA polymerase IIA associates with the promoter and, in the presence of ATP, is converted to RNA polymerase 110. Indeed, the formation of RNA polymerase 110, under these reaction conditions, appears to be dependent on the formation of a preinitiation complex (compare lanes 3 and 4 with lanes 6-8 in Fig. 20). The fact that RNA polymerase 110 does not appear to be formed in the absence of DNA under these conditions (Fig. 4C, lanes 6-8) suggests that phosphorylation of the C-terminal domain occurs after association of RNA polymerase with the preinitiation complex as opposed to RNA polymerase 110 selectively binding to the promoter. In order to determine the phosphorylation state of RNA polymerase subunit IIa throughout the course of transcription, preinitiation complexes were formed as described above and subsequently incubated in the presence of various combinations of nucleotides. Transcription reactions included either no nucleotides, ATP alone, ATP and CTP, or ATP, CTP, UTP, and 3'-0-methyl-GTP thereby resulting in the formation of either a preinitiated complex, an activated complex, an initiated complex, or an elongating complex, respectively (Van Dyke et al., 1988;Buratowski et al., 1989;Saltzman and Weinmann, 1989). Complexes were fractionated from free RNA polymerase II by gel filtration and analyzed as described above. In agreement with the previous experiment, incubation of RNA polymerase IIA with transcription factors and DNA results in the formation of preinitiation complexes containing RNA polymerase IIA (Fig. 3A, lanes 2-3). In the presence of ATP, RNA polymerase IIA is converted to RNA polymerase 110 (Fig. 3B, lanes 2-4). RNA polymerase II remains as 110 through initiation (Fig. 3C, lanes 2-3) and early elongation (Fig. 30, lanes 2-3). These results are in agreement with photoaffinity labeling experiments that demonstrate that the elongation phase of transcription is catalyzed by RNA polymerase 110 (Bartholomew et al., 1986;Cadena and Dahmus, 1987;Payne et al., 1989).
The observation that the conversion of RNA polymerase IIA to 110 occurs in the presence of ATP alone, implies that phosphorylation of the C-terminal domain precedes the formation of the first phosphodiester bond. The possibility exists, however, that trace amounts of contaminating nucleotides may be present in concentrations sufficient for initiation to occur. To test this possibility, the transcriptional activity of preinitiation complexes, formed in the presence of either ATP or dATP and purified by gel filtration, was determined in the presence of 0.08% Sarkosyl. If transcripts were initiated during the ATP incubation step, the purified transcription complexes would be stable in the presence of Sarkosyl, and completed transcripts would be formed upon the addition of remaining nucleotides (Hawley and Roeder, 1987). In order to monitor the formation of RNA polymerase 110 and to identify column fractions containing preinitiation complex, reactions were initiated by the addition of 32P-labeled RNA polymerase IIA. The experimental protocol is shown diagrammatically in Fig. 4. Results presented in Fig. 4A confirm that excluded fractions contain exclusively RNA polymerase 110 irrespective of whether the incubation contained ATP (lanes 1-7) or dATP (lanes 8-14). This is in agreement with previous results that show dATP can substitute for ATP in the conversion of RNA polymerase IIA to 110 .

Preinitiation
complexes were formed and fractionated by gel filtration, and column fractions were run on SDS-PAGE as described in the legend to Fig. 2. Panels A, B, C, and D contain column fractions from reactions incubated with DNA but without nucleotides (A), with ATP (B), with ATP and CTP (C), and with ATP, CTP, UTP, and 3'-0-meGTP (D). Samples were analyzed on a 5% polyacrylamide-SDS gel.
The two peak fractions containing preinitiation complexes were pooled and divided into two reactions. The first reaction was incubated with ATP, CTP, GTP, and UTP as described under "Experimental Procedures" for 5 min at which time Sarkosyl was added to a final concentration of 0.08% and the incubation continued for an additional 40 min (Fig. 4B, lanes  1 and 5). The second reaction was incubated for 5 min in the presence of Sarkosyl prior to the addition of nucleotides (Fig.  4B, lanes 2 and 6). As an additional control, the included fractions containing free RNA polymerase IIA were also assayed for transcriptional activity (Fig. 4B, lanes 3-4 and 7-8). Completed transcripts were formed only in reactions containing purified preinitiation complexes that were incubated with nucleotides prior to the addition of Sarkosyl (Fig. 4B,  lanes 1 and 5). The fact that no transcripts were formed when Sarkosyl was added prior to the addition of nucleotides establishes that transcripts were not initiated during the preincubation with ATP or dATP (Fig. 4B, lanes 2 and 6). These results therefore support the idea that phosphorylation of the C-terminal domain precedes the formation of an initiated complex. Preinitiation complexes (3X reaction or 120 ~1) were formed in the presence of either 700 pM ATP or dATP and 6 mM MgCl2 and purified on a 3-ml gel filtration column as described under "Experimental Procedures." Fractions of 150 ~1 were collected. Aliquots of 50 ~1 were analyzed by electrophoresis on 5% polyacrylamide-SDS gels (A). Fractions 8-12, 14, and 16 from each column were run in lanes I-7 (complexes formed in the presence of ATP) and in lanes 8-14 (complexes formed in the presence of dATP). The two peak fractions of each column containing preinitiation complexes (A, lanes 2-3 and 8-9) were pooled and assayed for transcriptional activity (B, lanes 1-2 and 5-6). Included fractions containing RNA polymerase (A, lanes 6 and 13) were also assayed for transcriptional activity (B, lanes 3-4 and 7-8). Lanes l-4 and 5-8 of panel B are from preincubation reactions that contained ATP and dATP, respectively. Reactions in lanes 1,3,5, and 7 were incubated with nucleotides prior to the addition of Sarkosyl, whereas reactions in lanes 2, 4, 6, and 8 were incubated with Sarkosyl prior to the addition of nucleotides.
sented in Figs. 2 through 4 establish that in the absence of ATP, RNA polymerase IIA is associated with the preinitiation complex, whereas in the presence of ATP, RNA polymerase 110 is associated with such complexes. If indeed the conversion of RNA polymerase IIA to 110 occurs after RNA polymerase IIA has associated with the preinitiation complex, the protein kinase that catalyzes the phosphorylation of the Cterminal domain of subunit IIa may be an integral component of such complexes. In order to test this possibility, preinitiation complexes were formed in the absence of ATP, purified by gel filtration, and assayed for RNA polymerase IIA to 110 conversion activity by incubation in the presence of ATP. Preinitiation complexes were formed by incubation of 32Plabeled RNA polymerase IIA with a reconstituted transcription extract and the AdZ-MLP as described above and purified by gel filtration. Column fractions were incubated in the presence of ATP, and the level of phosphorylation of subunit IIa/o was determined by SDS-PAGE (Fig. 5A). A major fraction of RNA polymerase IIA contained in excluded fractions was converted to RNA polymerase 110 upon incubation with ATP (Fig. 5, lanes 2-3). In an effort to determine if the recovery of protein kinase in fractions containing preinitiation complex is indeed dependent on preinitiation complex formation, transcription factors and DNA were incubated in the absence of RNA polymerase II and fractionated by gel filtration. Fractions were assayed for conversion activity, as described above, following the addition of ATP and 32Plabeled RNA polymerase IIA (Fig. 5B). The observation that incubation of RNA polymerase IIA and ATP in the presence of excluded fractions does not result in a significant shift in the electrophoretic mobility of subunit IIa, suggests that the association of protein kinase with the preinitiation complex is dependent on the presence of RNA polymerase II. The protein kinase responsible for the conversion of RNA polymerase IIA to 110 is also not recovered in the excluded fractions Preinitiation complexes were formed in the presence of 32P-labeled RNA polymerase IIA in the absence of ATP and purified by gel filtration as described under "Experimental Procedures" (panel A). Aliquots (50 ~1) of various fractions were incubated in the presence of 600 pM ATP for 15 min at 30 "C and analyzed by SDS-PAGE. An identical preincubation reaction, but lacking RNA polymerase II, was fractionated in parallel (panel B). Aliquots (50 ~1) of various fractions were incubated in the presence of 32P-labeled RNA polymerase IIA and ATP and analyzed as described above. Lanes l-8 correspond to fractions 8-12,14, 16, and 18. when RNA polymerase IIA is present but DNA is absent from the preincubation reaction (data not shown). These results are consistent with the idea that the protein kinase responsible for phosphorylation of the C-terminal domain is a com-ponent of TFIIE/F . Since TFIIE/F appears to interact directly with RNA polymerase II, the association of protein kinase with the complex would be expected to be dependent on the prior association of RNA polymerase II. The failure to observe the partial conversion of RNA polymerase IIA to 110 in included fractions is likely due to the dilution of protein kinase following chromatography.
Results presented above suggest that phosphorylation of RNA polymerase IIA may be facilitated by its association with the promoter. As a partial test of this idea, the concentration of ATP required for the phosphorylation of RNA polymerase IIA, free in solution and assembled into a preinitiation complex, was determined. Preinitiation complexes containing "'P-labeled RNA polymerase IIA were formed and purified as described above. Complexes were incubated in the presence of increasing concentrations of ATP, and the extent of RNA polymerase IIA to 110 conversion was determined. Results presented in Fig. 6A show that concentrations of ATP as low as 30 pM are sufficient for the conversion of RNA polymerase IIA to 110 when RNA polymerase IIA is associated to 110. Preinitiation complexes in a 180-~1 reaction containing casein kinase II "'P-labeled RNA polymerase IIA were formed in the absence of ATP and fractionated by gel filtration on a 6-ml column as described under "Experimental Procedures." Excluded peak fractions were pooled, and aliquots (40 ~1) were incubated at 30 "C for 15 min with 0, 10,30, 100, 300, and 1000 pM ATP (panel A, lanes 1-6, respectively) or GTP (panel B, lanes 1-6). Conversion reactions containing free "P-labeled RNA polymerase IIA and transcription factors but no DNA were incubated with 0, 10,30, 100,300, and 1000 FM ATP (panel C, lanes 2-7) or GTP (punel D, lanes Z-7). Samples were analyzed by SDS-PAGE as described above.
with the preinitiation complex. The effect of increasing concentrations of ATP on the conversion of free "P-labeled RNA polymerase IIA in the presence of the same transcription extract is shown in Fig. 6C. The conversion reaction in the absence of DNA was done with an otherwise complete reaction that was not fractionated by gel filtration. It is apparent from Fig. 6C that concentrations of ATP as high as 1 mM result in only the partial conversion of free RNA polymerase IIA to 110. The association of RNA polymerase IIA with the preinitiation complex, therefore, results in a two to three order of magnitude reduction in the ATP concentration required for conversion (compare Fig. 6A, lane 2 with Fig. 6C, lane 6). One possibility is that the increased efficiency of conversion of RNA polymerase IIA associated with transcription complexes results from the removal of an inhibitor or protein phosphatase during the purification of such complexes. This was tested by examining the effect of unfractionated transcription extract on the rate of conversion of RNA polymerase IIA in purified transcription complexes. The inclusion of extract did not have an appreciable effect on the rate of appearance of RNA polymerase 110 even though RNA polymerase IIA was detectable at low ATP concentrations (data not shown). The effect of increasing concentrations of GTP on the conversion of RNA polymerase IIA to 110, either associated with a transcription complex or free in solution, is shown in Fig. 6 B and D, respectively. In agreement with previous results , GTP can be utilized by the protein kinase that catalyzes the conversion of RNA polymerase IIA to 110. The apparent K, for GTP is, however, on the order of 100-fold higher than that of ATP (compare Fig. 6B, lane 6 with Fig. 6A, lane 2). Similar to the results with ATP, only partial conversion of free RNA polymerase IIA to 110 is observed in the presence of GTP (Fig. 40).
RNA Polymerases IIO, IIA, and IIB Have Similar ATP Requirements-Adenosine nucleoside triphosphate (ATP or dATP) containing a hydrolyzable p, y-phosphoanhydride bond is required for the activation of RNA polymerase II (Conaway and Conaway, 1988;Sawadogo and Roeder, 1984;Reinberg and Roeder, 1987). The nature of this energy dependence is not known. The ATP requiring step immediately precedes initiation and consequently appears to occur at about the same time as the conversion of RNA polymerase IIA to 110. The possibility therefore exists that the energy-dependent step is the phosphorylation of the C-terminal domain of subunit IIa. If the conversion of RNA polymerase IIA to 110 is the sole energy-requiring step, transcription catalyzed by RNA polymerase IIB, that lacks a C-terminal domain (Corden et al., 1985), should not be dependent on the presence of ATP.
Preinitiation complexes were formed as described above by incubation of purified RNA polymerase 110, IIA, or IIB with transcription factors, DNA template containing the Ad2-MLP, and increasing concentrations of ATP or dATP as indicated in the legend to Fig. 7. Transcription was initiated by the addition of AMP-PNP, CTP, GTP, and UTP and the amount of 560-nucleotide transcript produced was quantitated. AMP-PNP can be utilized as a substrate for elongation, but cannot satisfy the energy-requiring step. The results shown in Fig. 7 A   and the effect of a monoclonal antibody, that differentially reacts with RNA polymerases IIA and 110, on in. uitro transcription , also support the idea that RNA polymerase IIA interacts with the template-committed complex to form a preinitiation complex. Transcription reactions, initiated with RNA polymerase 110 (panel A), IIA (panel B), or IIB (panel C), were incubated with increasing concentrations of ATP or cL4TP at 25 "C for 15 min followed by a second incubation with DNA at 25 "C for 15 min. Transcription was initiated by the addition of AMP-PNP, [w~'P] CTP, -UTP, -GTP, and M&l2 as described under "Experimental Procedures." Each reaction contained 8 X 10e3 units of RNA polymerase II. The amount of 560 nucleotide transcript was quantitated as previously described , and the percent maximum activity was ~plotted as a function of the log of ATP (U) or dATP ([1--o) concentration. The 100% value corresponds to 0.77,0.74, and 0.86 x 10-l pmol of transcript produced for RNA polymerases 110, IIA, and IIB, respectively. The gel inset in each panel are reactions containing 0, 0.095,0.95,9.5,95, and 950 j.~cM ATP in lanes 1-6, respectively.
The inclusion of ATP in the transcription reaction resulted in the recovery of preinitiation complexes that contained almost exclusively RNA polymerase 110 (Figs. 20, 3B, and 4A). RNA polymerase II not associated with the preinitiation complex was, however, only partially phosphorylated. The exclusive localization of RNA polymerase 110 in the activated complex could result from either the preferential phosphorylation of RNA polymerase IIA in the preinitiation complex, relative to the free enzyme, or the preferential association of RNA polymerase 110 with the template-committed complex. The following observations support the idea that phosphorylation occurs subsequent to the association of RNA polymerase IIA with the promoter. The limited phosphorylation of unbound RNA polymerase II that occurs under standard transcription reaction conditions does not result in the formation of subunit 110 as determined by mobility shift in SDS gels (Fig. 20). Furthermore, results presented in Fig. 6 show that the phosphorylation of RNA polymerase IIA is stimulated by its association with the promoter.
IIA was consistently observed. The basis for this inhibition, or why RNA polymerases 110 and IIB do not exhibit similar sensitivities, is not known. The energy requirement could also be satisfied by the presence of dATP (only quantitative data shown), but not by GTP, UTP, CTP, and AMP-PNP, in agreement with previous findings (Sawadogo and Roeder, 1984;Reinberg and Roeder, 1987;Conaway and Conaway, 1988). These results suggest that, in addition to the ATP requirement for the phosphorylation of the C-terminal domain of subunit IIa, the initiation of transcription by RNA polymerase II must contain a second energy-requiring step.

DISCUSSION
The state of phosphorylation of RNA polymerase II is dependent on its position in the transcription cycle. Results The nucleotide sequence at the 5' end of the major late transcript is ACUCU. The formation of the first phosphodiester bond therefore requires the presence of both ATP and CTP. The observation that the phosphorylation of RNA polymerase II, bound in a preinitiation complex, requires only the presence of ATP or dATP indicates that phosphorylation precedes the formation of the first phosphodiester bond. Furthermore, the fact that transcription complexes formed in the presence of either ATP or dATP cannot synthesize completed transcripts in the presence of 0.08% Sarkosyl indicates that initiation did not take place during the ATP preincubation. The observation that preinitiation complexes formed in the presence of ATP contain almost exclusively RNA polymerase 110 suggests that the transcripts initiated upon addition of the remaining nucleotides are initiated by RNA polymerase 110. Quantitation of the amount of transcript produced indicates that about 10% of the RNA polymerase present in the preinitiation complex fraction is able to synthesize a completed transcript in the presence of Sarkosyl. It is. therefore,