THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1988 by The American Society for Biochemistry and Molecular Biology, Inc

Vol. 263, No. 6, Issue of February 25, pp. 2962-2968, 1988 Printed in U.S.A.

ATP Activates Transcription Initiation from Promoters by RNA Polymerase I1 in a Reversible Step Prior to RNA Synthesis*

(Received for publication, September 4, 1987)

Ronald C. Conaway and Joan Weliky Conaway4 From the DNAX Research Institute of Molecular and Cellular Biology, Palo Alto, California 94304

We have investigated the role that ATP plays in the synthesis of accurately initiated transcripts from the adenovirus 2 major late and mouse interleukin-3 pro- moters by a purified RNA polymerase I1 transcription system prepared from rat liver. The synthesis of 250- 330 nucleotide run-off transcripts and 4-9 nucleotide Sarkosyl-resistant transcription intermediates re- quires ATP both for RNA synthesis and for activation of the system prior to RNA synthesis. Activation spe- cifically requires an adenine nucleoside triphosphate containing a hydrolyzable &-y-phosphoanhydride bond. ATP, adenine-9-8-D-arabinofuranoside (araATP), and dATP are potent activators of transcription; they ac- tivate transcription to 50% of maximum at 2 PM. ATP analogs containing nonhydrolyzable @,-y-phosphoan- hydride bonds such as adeny1-5’-yl imidodephosphate, adenosine 5’-(fl,-y-methylene)triphosphate, and adeno- sine 5’-0-(thi0)triphosphate (ATP-yS) function effi- ciently in chain elongation, but do not activate tran- scription. Furthermore, ATP-yS is a potent, reversible inhibitor of ATP activation. 20 I.IM ATP-yS inhibits the synthesis of both full-length run-off transcripts and sarkosyl-resistant intermediates by 50% when the con- centration of ATP is 10 PM. ATP-yS inhibition can be overcome by high concentrations of ATP, dATP, araATP, or ddATP. Inhibition of the synthesis of Sar- kosyl-resistant transcription intermediates by ATP-yS is prevented by preincubation of the transcription en- zymes and DNA template with ATP and magnesium prior to the addition of ATP-yS and the remaining ribonucleoside triphosphates. Thus we argue that ATP activates the transcription system in a step prior to RNA synthesis.

Shortly after the discovery of methods for preparing tran- scriptionally active extracts from mammalian cells (1-3), it was shown that transcription initiation at several promoters by RNA polymerase I1 required ATP for some function be- sides RNA chain elongation (4-6). Bunick et al. (4) showed that AMP-PNP’ could replace ATP during transcription of the adenovirus VA RNAs by RNA polymerase 111; however, AMP-PNP could not replace ATP during the synthesis of run-off transcripts from the adenovirus 2 major late and e-

* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisernent” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ To whom reprint requests should be addressed DNAX Research

Alto, CA 94304. Institute of Molecular and Cellular Biology, 901 California Ave., Palo

The abbreviations used are: AMP-PNP, adenyl-5”yl imidodi- phosphate; araATP, adenine-9-6-D-arabinofuranoside 5”triphos- phate; dATP, 2’-deoxyadenosine 5”triphosphate; N1-oxide-ATP,

globin promoters by RNA polymerase 11. AMP-PNP, an ATP analog containing an imido group in place of the oxygen that bridges the P,y-phosphates, cannot be hydrolyzed to ADP and Pi by most phosphohydrolases (7), but it can be incorporated into growing RNA chains by RNA polymerase I1 (4,6). These results argued that ATP, and possibly the hydrolysis of ATP to ADP, was required for the synthesis of accurately initiated transcripts, but they did not clearly demonstrate whether ATP was required for the initiation of run-off transcripts, the elongation of run-off transcripts, or both. Subsequently, Sa- wadogo and Roeder (6) found that dATP could substitute for ATP if AMP-PNP was included in reactions as a substrate for RNA synthesis by RNA polymerase 11. More importantly, they showed that ATP or dATP was required for the synthesis of the first 9 nucleotides of transcripts initiated at the aden- ovirus 2 major late promoter.

Although it is clear that the initiation of RNA synthesis at several promoters by RNA polymerase I1 requires ATP, many questions remain concerning the nature of this requirement and the mechanism by which ATP facilitates transcription. Is ATP required prior to, during, or after the synthesis of phosphodiester bonds? Is ATP required once or at several stages of transcription initiation? Does ATP activate tran- scription? If so, is activation reversible or irreversible? What is the function of ATP in transcription initiation? Is it hy- drolyzed to provide energy, to serve as a phosphate donor in phosphorylation, or both; or, is it merely bound by compo- nent(s) of the transcription system? Finally, which transcrip- tion factors mediate ATP dependence?

We have recently described the preparation of a purified RNA polymerase I1 transcription system from rat liver (8). In this report, we show that ATPyS is a powerful inhibitor of the ATP-requiring step in transcription initiation, and we use ATPyS to obtain evidence supporting specific answers to several of these questions.

EXPERIMENTAL PROCEDURES

Materials-Unlabeled ultrapure ribonucleoside 5”triphosphates and 2’-deoxynucleoside 5’-triphosphates were purchased from Phar- macia LKB Biotechnology Inc. AMP-PCP, ddATP, dITP, 6-ATP, 5’- ADP, 5’-AMP, ApC, ApA, and ApG were also from Pharmacia. GMP- PNP, GTPrS, araATP, N’-oxide-ATP, w-hexyl-ATP, XTP, and 3’- AMP were from Sigma. AMP-PNP was obtained from Sigma or Pharmacia, and ATPrS was from Pharmacia or Boehringer Mann-

adenosine N’-oxide 5”triphospbate; t-dATP, 1-N6-etheno-2”deox- yadenosine 5’-triphosphate; XTP, xanthosine 5’-triphosphate; dITP, 2’-deoxyinosine 5’-triphosphate; N6-hexyl-ATP, N6-[(6-amino- hexyl)carbamoylmethyl]adenosine 5”triphosphate; ATPyS, adeno- sine 5’-O-(thio)triphosphate; GTP+, guanosine 5’-0-(thio)tri- phosphate; AMP-PCP, adenosine 5’-(@,y-methylene)triphosphate; GMP-PNP, guanyl-5“yl imidodiphosphate; ApC, adenyl-3’-5’-yl cy- tidine; ApA, adenyl-3’-5’-yl adenosine; ApG, adenyl-3’-5’-yl guano- sine; 5’-ADP, adenosine 5”diphosphate; 5’-AMP, adenosine 5’-mon- ophosphate; 3’-AMP, adenosine 3’-monophosphate.

2962

ATP Activation of Transcription Initiation by RNA Polymerase 11 2963

heim. [a-32P]CTP (400 Ci/mmol), [cY-~'P]GTP (400 Ci/mmol), and [cY-~'P]ATP (400 Ci/mmol) were obtained from Amersham Corp. Bovine serum albumin (reagent grade) was from Miles Biochemicals. Human placental ribonuclease inhibitor was purchased from Be- thesda &search Laboratories or Promega Biotech. Polyvinyl alcohol, type 11, and Sarkosyl (N-lauroylsarcosine) were from Sigma.

Enzymes-RNA polymerase 11, transcription protein a (Fraction V), and fractions B' and D were purified from rat liver as previously described (8).

DNA Templates-Plasmid pmIL-3 (8,9,12), containing the mouse interleukin-3 promoter, was digested with NdeI a t a unique site 330 nucleotides downstream from the start site of transcription. Plasmid pDNAdML was constructed by cloning a synthetic oligonucleotide containing the adenovirus 2 major late promoter sequence from -50 to +10 (10, 13) into the KpnI and XbaI sites of pUC-18. Plasmid pDNAdML was digested with NdeI at a unique site 250 nucleotides downstream from the start site of transcription.

Assay of Run-off Transcription-Except as indicated in the figure legends, assays were performed as described (8) with 0.1 pg of NdeI digested pmIL-3 or pDNAdML as template, 3 pg of fraction D, 1 pg of fraction B', 5 ng of transcription protein a, and 0.01 units of RNA polymerase 11.

RESULTS

ATP Is Required for the Synthesis of Full-length Run-off Transcripts by the Rat Liver Transcription System-Template DNA containing the adenovirus 2 major late promoter was transcribed by a purified rat liver enzyme system that sup- ports accurate initiation by RNA polymerase I1 (8). The synthesis of full-length run-off transcripts was optimal when reaction mixtures contained 5 p~ ATP, 10 pM CTP, 10 pM UTP, and 10 p~ GTP (Fig. 1, lane 1) . If GTP was replaced with GMP-PNP, run-off transcripts were still synthesized (data not shown). In contrast, when ATP was replaced with

+AMPPNP -AMPPNP I d-

AdML- 4

C

4

1 2 3 4 5 6 7 8 9 FIG. 1. The synthesis of run-off transcripts depends on ATP

or ATP analogs. Transcription reactions were performed as de- scribed under "Experimental Procedures." The DNA template was pDNAdML. In addition to 10 p~ UTP, 10 p~ GTP, 10 p~ [a-3ZP] CTP (25 Ci/mmol), and 7 mM MgCIZ, reactions contained 5 pM ATP (lanes 1 and 3), 5 p~ dATP (lanes 4 and 7), 5 PM araATP (lanes 5 and 8), or 40 p~ ddATP (lanes 6 and 9) . Reactions were performed with (lanes 2-6) or without (lanes 7-9) 100 p~ AMP-PNP.

AMP-PNP, no run-off transcription was observed (Fig. 1, lane 2), even though AMP-PNP is a substrate for chain elongation by RNA polymerase I1 (4, 5). AMP-PNP did not inhibit transcription, however, since run-off transcripts could be synthesized when ATP was added to reactions that con- tained AMP-PNP (Fig. 1, lane 3). In addition, several analogs of ATP, including dATP, araATP, and ddATP, could restore transcription when AMP-PNP, CTP, GTP, and UTP were present as substrates for RNA synthesis by RNA polymerase I1 (Fig. 1, lanes 4-6); but they could not restore transcription when AMP-PNP was omitted from reactions (Fig. 1, lanes 7- 9). Similar results were obtained when a plasmid containing the mouse interleukin 3 promoter was used as template (data not shown). Thus, in the rat liver transcription system, ATP is not only required for chain elongation, but it is also required for some additional function.

In order to determine what features of the ATP molecule are recognized by the transcription system, we tested a variety of ATP analogs for their ability to facilitate the synthesis of run-off transcripts in reactions containing AMP-PNP, CTP, UTP, and GTP (Table I). Of the analogs tested, only nucle- oside triphosphates containing intact j3,y-phosphates and an adenine (6-aminopurine) base were active. Nucleotides with missing or modified j3,y-phosphates (AMP-PNP, AMP-PCP, ATPyS, 5'-ADP, 5'-AMP, and 3'-AMP) were inactive. More- over, nucleotides lacking a 6-aminopurine base (CTP, GTP, UTP, dCTP, dGTP, dUTP, deoxyinosine triphosphate, and xanthosine triphosphate) were also inactive. Modifications to the 6-aminopurine base could be tolerated, however, since t- dATP, N'-oxide-ATP, and N'-hexyl-ATP were all active. Although the structure of the triphosphate and the base had to be preserved, the pentose sugar could be modified without abolishing transcription, since araATP, dATP, and ddATP were all active.

ATP Is Required for the Synthesis of Sarkosyl-resistant Transcription Intermediates-The synthesis of full-length run-off transcripts is a complex process involving both initi- ation and subsequent elongation of RNA chains by RNA polymerase 11. Therefore, we could not use the standard run- off transcription assay to determine whether ATP is required prior to, during, or after initiation in the rat liver system. In HeLa cell transcription systems, it has been shown that initiation by RNA polymerase I1 is inhibited by concentra-

TABLE I Effect of ATP analogs on run-off transcription

Reactions were performed as described under "Experimental Pro- cedures." The DNA template was pDNAdML. In addition to the nucleotides indicated in the table, reaction mixtures contained 10 pM UTP, 10 pM GTP, 10 p~ [a-"PICTP (25 Ci/mmol), and 60 pM AMP- PNP.

Nucleotide Concentration transcription Run-off

PM % m i m u m ATP 2 50 dATP 2 50 araATP 2 50 ddATP 20 50

N'-Oxide-ATP 2 50 N'-Hexyl-ATP 5 50 dITP 100 <1 XTP 100 <1

d A T P 15 50

AMP-PNP 100 <1 AMP-PCP 100 <1 ATPyS 100 <1 5"ADP 50 <1 5'-AMP 150 <1 3"AMP 150 <1

2964 ATP Activation of Transcription Initiation by RNA Polymerase II

tions of Sarkosyl greater than 0.05- 0.1%, while elongation of previously initiated RNA chains is unaffected. Furthermore, it has been suggested that one or at most a few phosphodiester bonds need to be synthesized in order to render further chain elongation resistant to inhibition by Sarkosyl (11, 12). Thus, Sarkosyl appears to separate run-off RNA synthesis into two discrete steps: 1) the initiation and synthesis of a short oligoribonucleotide in a step that can be inhibited by Sarkosyl and 2) the completion of a full-length run-off transcript by RNA polymerase I1 in a step that is resistant to inhibition by Sarkosyl. We therefore sought to characterize the effect of Sarkosyl on transcription in the rat liver system and to ask whether ATP is required during the synthesis of Sarkosyl- resistant transcription intermediates.

The synthesis of full-length run-off transcripts from either the adenovirus 2 major late or mouse interleukin-3 promoters was completely abolished when 0.25% Sarkosyl was added to reaction mixtures prior to the addition of ribonucleoside tri- phosphates (Fig. 2, compare lane 2 to lane 1 and lane 5 to lane 6). In contrast, when Sarkosyl was added to reactions 1 min after the addition of the four ribonucleoside triphosphates and magnesium (at which time no full-length run-off tran- scripts have been synthesized, Fig. 2, lunes 4 and 7), run-off transcripts were synthesized from both promoters (Fig. 2, lanes 3 and 8). In the presence of Sarkosyl, synthesis of full- length run-off transcripts proceeded at a rate 25-40% of maximum. Full-length transcripts first appeared within 10 min, and the reaction was complete in 30-40 min. The maxi- mum amount of full-length run-off transcript synthesized in the presence of Sarkosyl was 10-50% of that synthesized in the absence of Sarkosyl (data not shown). Therefore, Sarkosyl prevents initiation of transcription at promoters in the rat liver transcription system. When Sarkosyl is added after initiation has occurred, however, it does not prevent the elongation of previously initiated transcription intermediates.

Ribonucleoside triphosphates required for the formation of

IL-3 -

AdML-

FIG. 2. Sarkosyl inhibits initiation but not elongation of run-off transcripts. Reactions were performed as described under “Experimental Procedures.” The DNA template was either pDNAdML (lanes 1-4) or pmIL-3 (lanes 5-8). 0.25% Sarkosyl was added immediately prior to (lanes 2 and 5) or 1 min after (lanes 3 and 8) addition of 10 p~ ATP, 10 pM GTP, 10 pM UTP, 10 p M [a- “‘PICTP (25 Ci/mmol), and 7 mM MgCI,. No Sarkosyl was added to reactions in lanes 1, 4, 6, and 7. Reactions in lanes 4 and 7 were stopped 1 min after addition of labeled ribonucleoside triDhosDhates.

Sarkosyl-resistant transcription intermediates were deter- mined. When plasmid pDNAdML, containing the adenovirus 2 major late promoter, was used as template, Sarkosyl-resist- ant transcription intermediates could be synthesized only when magnesium and either all four ribonucleoside triphos- phates or ATP, CTP, and UTP were included in reaction mixtures prior to the addition of Sarkosyl (Fig. 3A). When plasmid pmIL-3, containing the mouse interleukin-3 pro- moter, was used as template, Sarkosyl-resistant transcription intermediates could be synthesized only when magnesium and either all four ribonucleoside triphosphates or ATP, GTP, and CTP were included in reaction mixtures prior to the addition of Sarkosyl (Fig. 3B). Thus, synthesis of Sarkosyl- resistant transcription intermediates requires only those ri- bonucleoside triphosphates necessary for the synthesis of a 4-9 nucleotide RNA (see Fig. 3, A and B, lower panels).

ATP is required for formation of Sarkosyl-resistant tran- scription intermediates. When ATP was replaced with AMP- PNP prior to Sarkosyl addition, no full-length run-off tran- scripts were observed, even though ATP was added to reaction mixtures after Sarkosyl addition (Fig. 4, lane 2). RNA syn- thesis could be restored to the same level observed with ATP alone (Fig. 4, lane 1 ) if either dATP (Fig. 4, lane 3) or ATP (Fig. 4, lane 4 ) was included with AMP-PNP and the other ribonucleoside triphosphates prior to Sarkosyl addition. Sim- ilar results were obtained when either araATP or ddATP was included with AMP-PNP and the other ribonucleoside tri- phosphates prior to Sarkosyl addition (data not shown). ATP, araATP, and dATP were most active and gave 50% of maxi- mum run-off synthesis at 2 pM and more then 95% of maxi- mum run-off synthesis at 5 pM. These results suggest strongly that ATP or an ATP analog with a hydrolyzable B,y-phos- phoanhydride bond is needed for the synthesis of at least the first 4-9 nucleotides of each run-off RNA.

RNA polymerase I1 can use the dinucleoside monophos- phate, ApC, as a substrate in the initiation of run-off tran- scripts at the adenovirus 2 major late promoter (13). We asked

A B G G U 0 U G U G U G C C G U C C U C C G U C C U A A A A A A A A A A A A A A A A

AdML - - 0

r rr ACUCUCUUCCUCAUG ... AGAACCCCUUG

FIG. 3. Ribonucleoside triphosphates required for synthesis of Sarkosyl-resistant transcription intermediates. Reactions were performed as described under “Experimental Procedures.” The DNA template was either pDNAdML (panel A ) or pmIL-3 (panel B ) . 10 p~ GTP, 10 p M UTP, 10 pM CTP, and 10 p~ [a-”’PJATP (25 Ci/ mmol) were included in reaction mixtures as indicated in the figure. After 1 min, 0.25% Sarkosyl was added to each reaction. After an additional 30 s, reactions were made to 40 pM GTP, 40 pM UTP, and 40 UM CTP: and reaction mixtures were incubated for 30 min. IL-3,

IL-3, interleukin-3. ”

intheukin-3.

ATP Activation of Transcription Initiation by RNA Polymerase II 2965

AMPPNP - -I- + + ATP -I- - - i-

dATP - - + - ori- 1.) a a’- 1

IL-3- i I

I

I

1 2 3 4 FIG. 4. ATP is required for the synthesis of Sarkosyl-re-

sistant transcription intermediates. Reactions were performed as described under “Experimental Procedures.” The DNA template was pmIL-3. Reaction mixtures contained 10 p~ GTP, 10 pM [a-”PICTP (25 Ci/mmol), and 10 p~ ATP ( l a n e I ) , 60 p~ AMP-PNP ( l a n e 2), 60 pM AMP-PNP, and 10 pM dATP ( l a n e 3), or 60 pM AMP-PNP and 10 p~ ATP ( l a n e 4 ) . After 1 min, 0.25% Sarkosyl, 10 pM UTP. and an additional 10 p~ ATP were added and reaction mixtures were incubated for 30 min. ZL-3, interleukin-e.

A. araATP - - -

A p C + + + + + + + + +

U T P - + + - + + C T P - - + - - + . . - ori

a -AdML

”_ . 1 2 3 4 5 6

B.

araA - + - + ApC - - + +

w 5 -ori

0 -AdML

1 2 3 4

FIG. 5. AraATP is required for synthesis of Sarkosyl-re- sistant transcription intermediates when transcription is in- itiated with ApC. Reactions were performed as described under “Experimental Procedures.” The DNA template was pDNAdML. Reactions mixtures in p a n e l A were incubated with 5 p~ araATP (lanes 4 4 3 , 10 p~ UTP (lanes 2, 3, 5, and 6 ) , 10 p~ CTP (lanes 3 and 6 ) , and 100 p~ ApC (lanes 1-6). After 1 min, 0.25% Sarkosyl and 10 p~ [CY-~’P]GTP (25 Ci/mmol) were added, and ATP, UTP, and CTP were brought to 10 p~ each. In addition to 10 p~ UTP and 10 pM [a-32P]CTP (25 Ci/mmol), reaction mixtures in panel B were incubated for 1 min with 5 p~ araATP (lanes 2 and 4 ) and 100 p~ ApC (lanes 3 and 4 ) prior to the addition of 0.25% Sarkosyl, 10 pM GTP, and 10 p~ ATP.

whether the requirement for ATP or an ATP analog in the synthesis of Sarkosyl-resistant transcription intermediates could be eliminated if the first phosphodiester bond was

supplied pre-formed as ApC. Consistent with our previous results, synthesis of Sarkosyl-resistant intermediates required ApC, CTP, and UTP (Fig. 5); neither ApA nor ApG could replace ApC (data not shown). Furthermore, transcription was dependent on araATP (Fig. 5); thus ATP or an ATP analog was still required for the synthesis of Sarkosyl-resist- ant transcription intermediates even if the first phosphodies- ter bond was supplied pre-formed.

ATPyS Is a Potent, Reversible Inhibitor of the Synthesis of Both Full-length Run-off Transcripts and Sarkosyl-resistant Transcription Intermediates-ATPyS is a potent inhibitor of run-off transcription from mouse interleukin-3 and adenovi- rus 2 major late promoters (Fig. 6, compare lane 1 to lane 3 and compare lane 4 to lane 6). When the concentration of ATP was 10 phi, the synthesis of run-off transcripts from both promoters was inhibited by 50% when the concentration of ATPyS was 20 phi, and it was inhibited by more than 95% when the concentration of ATPyS was 100 phi. GTPyS at 100 phi, however, had no effect on run-off transcription (data not shown). Inhibition by ATPyS could be reversed by addi- tion of high concentrations of ATP, dATP, araATP (Fig. 7), or ddATP (data not shown) to reaction mixtures that had been preincubated with ATPyS. Inhibition by ATPyS could not be reversed by AMP-PNP (see Fig. 6, lanes 3 and 6). ATPyS did not inhibit elongation by RNA polymerase 11. Instead, it could replace ATP or AMP-PNP as a substrate for RNA polymerase I1 if dATP, araATP, or ddATP were in- cluded a t concentrations sufficient to overcome inhibition of transcription by the ATPyS in the reactions (Fig. 8). These results suggest that ATPyS inhibits the synthesis of full- length run-off transcripts not by preventing chain elongation but rather by preventing ATP from carrying out its additional role in the synthesis of accurately initiated transcripts.

ATP and Magnesium Can Activate Transcription in the Absence of Other Ribonucleoside Triphosphates-RNA polym- erase 11, factors, and template DNA containing the adenovirus 2 major late promoter were preincubated with ATP and

AMPPNP -I- + + ATPyS - - +

ATP - i- P- -ori

+ + + - - + + - +

-ori

- I L-3

4 5 6 1 2 3

-AdML

FIG. 6. ATPyS inhibits synthesis of run-off transcripts. Re- actions were performed as described under “Experimental Proce- dures.” The DNA template was either pmIL-3 (lanes 1-3) or pDNAdML (lanes 4-6) . In addition to 10 p~ UTP, 10 p~ GTP, 60 pM AMP-PNP, and 10 pM [a-”PICTP (25 Ci/mmol), reaction mix- tures contained 10 p~ ATP (lanes I, 3 , 4 , and 6 ) and 100 p~ ATPrS (lanes 3 and 6 ) . ZL-3, interleukin-3.

2966 ATP Activation of Transcription Initiation by RNA Polymerase II

ATP dATP araATP 5 5 100 5 100 5 100pM -”

ATPyS - + + + + + + ori - *

AdML -

”_

1 2 3 4 5 6 7 ~ ~~

FIG. 7. Inhibition by ATPrS is reversible. Reactions were performed as described under “Experimental Procedures.” The DNA template was pDNAdML. 100 p~ ATP+ was added to some reac- tions (lanes 2-7) 30 s prior to the addition of 60 p~ AMP-PNP, 10 pM UTP, 10 p~ GTP, 10 p~ [a-”PICTP (25 Ci/mmol), and the indicated concentrations of ATP (lanes 1-3), dATP (lanes 4 and 5). and araATP (lanes 6 and 7). Reactions were incubated for 30 min.

ATPyS+ + - + - + - dATP- + - -

araATP - - - -4- i- - - ddATP - - - - - + +

- -

\ * O o i - ori

I !

I I

1 2 3 4 5 6 7 FIG. 8. ATPrS is a substrate for RNA synthesis by RNA

polymerase 11. Reactions were performed as described under “Ex- perimental Procedures.” The DNA template was pDNAdML. In addition to 10 p~ UTP, 10 p~ GTP, and 10 p~ [a-32P]CTP (25 Ci/ mmol), reaction mixtures contained 30 p~ ATPrS (lanes I , 2,4, and 6), 100 p~ dATP (lanes 2 and 3 ) . 100 p~ araATP (lanes 4 and 5), and 225 p~ ddATP (lanes 6 and 7) .

magnesium for 1 min prior to the addition of ATPyS, CTP, and UTP. If Sarkosyl and GTP were added after an addition 1-min incubation, full-length run-off transcripts were synthe- sized at a level close to that obtained in reactions containing no ATPyS (Fig. 9, compare lanes 1 and 3). When ATPyS

ATPYS at t =

ATP ATP dATP araATP

- 0’ 1‘ - 0’ 1‘ - 0’ 1’ - 0’ 1’ - - ”

AdML- 0 o

11-3- 0

11-3- 0 o

1 2 3 4 5 6 7 8 9 10 11 12

FIG. 9. ATP or ATP analogs activate the transcription sys- tem prior to the addition of the ribonucleoside triphosphates needed for RNA synthesis. Reactions were performed as described under “Experimental Procedures.” The DNA template was either pDNAdML (lanes 1-3) or pmIL-3 (lanes 4-12). 7 mM magnesium chloride and 3 pM ATP (lanes I d ) , 3 pM dATP (lunes 7-9), or 3 pM araATP (lanes 10-12) were added to reaction mixtures. After incu- bation for 1 min, 10 p~ GTP, 10 p~ UTP, and 10 pM [cY-”~P]CTP (25 Ci/mmol) (lanes 1-6) or 10 pM GTP, 10 pM UTP, 10 pM [a-”’P] CTP (25 Ci/mmol), and 60 p~ AMP-PNP (lanes 7-12) were added. After a further 1-min incubation, 0.25% Sarkosyl was added. 30 s later, ATP, GTP, and UTP were brought to 40 p~ each and reactions were incubated for 30 min. 100 pM ATPrS was included in reaction mixtures either with the magnesium chloride and ATP, MTP, or araATP ( t = 0, lanes 2, 5, 8, and 11) or 1 min after the addition of magnesium chloride and ATP, dATP, or araATP ( t = 1, lanes 3,6,9, and 12). ZL-3, interleukin-3.

was added at the beginning of the preincubation, however, no RNA synthesis occurred (Fig. 9, lane 2) . The same results were obtained when template DNA containing the mouse interleukin-3 promoter was used as template. ATP, magne- sium, template DNA, RNA polymerase 11, and factors were preincubated for 1 min prior to the addition of ATPyS, GTP, and CTP. If Sarkosyl and UTP were added after an additional 1-min incubation, full-length run-off transcripts were synthe- sized at a level close to that obtained in reactions containing no ATPyS (Fig. 9, lanes 4 and 6). When ATPyS was included in the preincubation, no RNA synthesis occurred (Fig. 9, lune 5). Similar results were obtained when magnesium and dATP (Fig. 9, lanes 7-9), araATP (Fig. 9, lunes IO-12), or ddATP (data not shown) were substituted for ATP during the initial preincubation. Thus, magnesium and ATP, dATP, araATP, or ddATP, in the absence of any other ribonucleoside tri- phosphates, can activate the transcription system; while the system is activated, RNA synthesis cannot be inhibited by ATPyS.

To ask whether ATP activation of the transcription system is reversible, we carried out the following experiment. RNA polymerase 11, factors, and template DNA containing the adenovirus major late promoter were incubated for 1 min with magnesium and ATP; then ATPyS was added. UTP and CTP were added a t various times after the addition of ATPyS; Sarkosyl and GTP were added 1 min later. The results of this experiment are shown in Fig. 10. The transcription system was rapidly inactivated after addition of ATPyS. In the presence of 100 PM ATPyS and 5 PM ATP, the half-life of the activated system was approximately 40 s.

DISCUSSION

It has been known for some time that ATP plays multiple roles during the initiation of RNA synthesis at promoters by

ATP Activation of Transcription Initiation by RNA Polymerase 11 2967

0' 1' 1.5' 2' 3' 5' ori- 7 w - ---

AdML- 0 -

"

1 2 3 4 5 6 FIG. 10. ATP activation of transcription is reversible. Re-

actions were performed as described under "Experimental Proce- dures." The DNA template was pDNAdML. 7 mM magnesium chlo- ride and 5 pM ATP were added to reaction mixtures. After incubation for 1 min, 100 p M ATPyS was added. 10 p~ UTP and [a-"PICTP (25 Ci/mmol) were added to reaction mixtures along with ATPyS ( l a n e I ) or at the indicated times after addition of ATPyS (lanes 2- 6). 1 min after the addition of CTP and UTP, Sarkosyl was added to 0.25% and GTP was brought to 10 p ~ .

human RNA polymerase I1 (4-6). In these studies, ATP was shown to be required as a substrate for RNA polymerase I1 in RNA synthesis and for the facilitation of transcription initi- ation by some unknown mechanism. We have confirmed and extended these observations with a purified RNA polymerase I1 transcription system from rat liver. In the process, we have made findings that shed light on how ATP facilitates tran- scription initiation by RNA polymerase 11.

ATP or an ATP analog is required for the synthesis of both 250-330-nucleotide run-off transcripts and 4-9-nucleotide Sarkosyl-resistant transcription intermediates from the ad- enovirus 2 major late and mouse interleukin-3 promoters. ATP analogs must contain a hydrolyzable P,y-phosphoanhy- dride bond. Neither AMP-PNP, AMP-PCP, nor ATPyS will replace ATP in transcription initiation even though each is a substrate for RNA polymerase I1 in RNA synthesis. Active ATP analogs must also contain an intact adenine (6-amino- purine) base. Of the purines tested, those lacking the 6-amino group did not replace ATP; however, purines containing mod- ifications of the adenine base such as e-dATP, N '-oxide-ATP, and N6-hexyl-ATP are active. Finally, the pentose sugar can be modified without abolishing transcription; in addition to dATP, araATP and ddATP can replace ATP.

ATPyS is a potent and reversible inhibitor of the ATP requiring step in the synthesis of both full-length run-off transcripts and Sarkosyl-resistant transcription intermedi- ates. ATPyS inhibits transcription facilitated by ATP, dATP, araATP, or ddATP; however, inhibition can be overcome by the addition of these analogs a t appropriately high concentra- tions. When dATP, araATP, or ddATP are present at high concentration, ATPyS is a substrate for RNA polymerase I1 in RNA synthesis.

Preincubation of magnesium and ATP, dATP, or araATP with transcription enzymes and DNA template prior to the

addition of ATPyS and the remaining ribonucleoside triphos- phates was sufficient to prevent inhibition of transcription initiation by ATPyS. If there was a delay of more than 1-2 min between addition of ATPyS and the remaining ribonu- cleoside triphosphates, the level of run-off transcription was significantly decreased. We therefore argue that ATP and magnesium reversibly activate the transcription system prior to the initiation of RNA synthesis. This notion is supported by the observation that ATP analogs that are not used as substrates for RNA synthesis by RNA polymerase I1 (dATP, araATP, ddATP) also render run-off transcription resistant to inhibition by ATPyS.

The simplest explanation for how ATP activates the tran- scription system prior to RNA synthesis is that either ATP binding or ATP hydrolysis pushes the system into a confor- mation that can rapidly initiate transcription. Once activated, the system will initiate transcripts if the remaining ribonucle- oside triphosphates are available. On the other hand, the activated system can be inactivated by a brief incubation with ATPyS in the absence of ribonucleoside triphosphates. Thus, in the presence of ATP and magnesium, there is most likely an equilibrium between active and inactive conformations of the transcription system.

Since an ATP analog was still required to activate the transcription system when the dinucleotide ApC was used to initiate RNA synthesis from the adenovirus 2 major late promoter, the ATP analog is probably not directly involved in the formation of the first phosphodiester bond. A separate ATP activation event is not required for the synthesis of each phosphodiester bond of the nascent transcript, since the ac- tivated transcription system could initiate and synthesize 4- 9-nucleotide long Sarkosyl-resistant intermediates even if ATPyS was added along with the ribonucleoside triphos- phates needed for RNA synthesis.

The experiments reported here do not tell whether ATP is hydrolyzed or merely bound by component(s) of the transcrip- tion system. If ATP hydrolysis is required to activate the system, we cannot tell whether ATP is used as a source of energy for some step in initiation, such as unwinding of the DNA, or whether ATP serves as a phosphate donor in a phosphorylation step. Because ATP activation of the tran- scription system is reversible in the presence of ATPyS, it is less likely that the transcription system is activated by phos- phorylation; we cannot, however, rule out the possibility that a phosphatase dephosphorylates and inactivates some com- ponent of the transcription system. Finally, we cannot deter- mine whether the transcription factors that mediate this ATP activation are directly involved in initiation of RNA synthesis. Because the assay for ATP activation is measurement of RNA synthesis, it is not possible to determine whether specific transcription factors are required only for activation, only for RNA synthesis, or for both. An assay capable of measuring ATP activation directly is needed; however, the development of such an assay could prove difficult because it appears that the activation step occurs close in time to RNA chain initia- tion.

Acknowledgments-We thank Doug Nomura for providing pDNAdML and Paula Wolf for expert technical assistance. We also thank Drs. Roger Kornberg, Gerard Zurawski, and J. Allan Waitz for critical readings of the manuscript and Jan Owens for help in pre- paring the manuscript.

REFERENCES 1. Weil, P. A., Luse, D. S. , Segall, J., and Roeder, R. G. (1979) Cell

2. Manley, J. L., Fire, A., Cano, A., Sharp, P. A., and Gefter, M. L. 18,469-484

(1980) Proc. Natl. Acad. Sci. U. S. A. 77,3855-3859

2968 ATP Activation of Transcription Initiation by RNA Polymerase II 3. Dignam, J. D., Lebovitz, R. M., and Roeder, R. G. (1983) Nucleic Chem. 262,8293-8297

4. Bunick, D., Zandomeni, R., Ackerman, S., and Weinmann, R. Acad. Sci. U. S. A. 8 2 , 316-320

5. Ernst, H., Filipowicz, W., and Shatkin, A. J. (1983) Mol. Cell. bon, P. (1982) Proc. Natl. Acad. Sci. U. S. A. 79 , 7132-7136

6. Sawadogo, M., and Roeder, R. G. (1984) J. Bid. Chem. 2 5 9 , 8172

7. Yount, R. G. (1975) Adu. Enzymol. Reht. Areas Mol. Bid. 4 3 , l - 3452-3461

8. Conaway, J. W., Bond, M. W., and Conaway, R. C. (1987) J. Biol. 259,2517-2525

Acids Res. 11, 1475-1489 9. Miyatake, S., Yokota, T., Lee, F., and Arai, K. (1985) Proc. Natl.

(1982) Cell 29,877-886 10. Hen, R., Sassone-Corsi, P., Corden, J., Gaub, M. P., and Cham-

Biol.3,2172-2179 11. Hawley, D. K., Roeder, R. G. (1985) J. Biol. Chem. 260 , 8163-

5321-5326 12. Hawley, D. K., and Roeder, R. G. (1987) J. Biol. Chem. 2 6 2 ,

56 13. Samuels, M., Fire, A., and Sharp, P. A. (1984) J. Biol. Chem.