topology of the product binding site in rna polymerase revealed

THE JOURNAL OF BIOLOGICAL CHEMISTFY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 269, No. 50, Issue of December 16, pp. 3170141705, 1994 Printed in U.S.A.

Topology of the Product Binding Site in RNA Polymerase Revealed by Transcript Slippage at the Phage A P, Promoter*

(Received for publication, May 16, 1994, and in revised form, August 31, 1994)

Konstantin SeverinovS and Alex GoldfarbO From the Public Health Research Institute, New York, New York 10016

In the presence of transcription substrates ATP, CTP, and UTP, a stable ternary complex containing tetranucleotide AUCA is formed on the phage A P, promoter (starting sequence C-&C-,A+,U+,C+.&G+&. We show that in the absence of GTP or at undersaturating GTP concentrations the AUCA transcript synthesized at the +1 to +4 segment slips back by 3 nucleotides and is stabilized in the ternary complex in such a way that only its 2 3’-proximal bases remain paired to the -l/+l positions of the template DNA. The slipped transcript can be extended in a template-directed manner into longer chains that can be cleaved by the GreA or GreB proteins at the +1/+2 junction. The slipped stabilized tetranucleotide delineates the “tight product binding site” of RNA polymerase responsible for stable holding of the transcript in the ternary transcription complex. The results suggest that the tight product binding site encompasses the locality within the complex where the nascent transcript detaches from the template strand of DNA.

Initiation of RNA synthesis by Escherichia coli RNA polymerase usually includes the stage of abortive initiation (1-3). At this stage, the polymerase-promoter complex catalyzes synthesis of short RNAs (2-9 nucleotides in length) that are rapidly released. After each short transcript is aborted, RNA polymerase remains bound to the promoter and the idle cycle repeats (2, 3). Transcription complexes that escape into elongation un- dergo a major structural rearrangement. This rearrangement usually occurs when the nascent RNA reaches the critical length of 9-11 nucleotides and is accompanied by release of the u subunit, stabilization of RNA in the complex, and commence- ment of processive elongation by the catalytic core component (a,PP’) (4-7). The stable holding of RNA in the rearranged ternary complex is of paramount biological significance since it ensures the remarkable processivity of RNA polymerase. Cur- rent models of elongation mechanism (8-13) attribute stable holding to a specific tight product binding site (TBS)’ in RNA polymerase located at a certain distance upstream from the RNAgrowing tip. Presumably this site gets “filled with RNA at the moment of promoter clearance.

When the sequence of promoter DNA near the transcription start point contains a tract of identical residues (e.g. oligo(dA)) or a short direct repeat, slippage of the nascent transcript within the promoter complex may occur (14-17). Recently, a

* This work was supported by National Institutes of Health Grant GM49242. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

nue, New York, NY 10021. t Present address: Box 224, Rockefeller University, 1230 York Ave-

8 To whom correspondence should be addressed: Public Health Re- search Inst., 455 First Ave., New York, NY 10016. “el.: 212-578-0839;

The abbreviations used are: TBS, tight product binding site; LBS, Fax: 212-578-0804.

loose product binding site.

stable ternary complex, which contained (r factor and carried tetrameric RNA CACC was described on the rrnB P1 promoter (17). The stable complex apparently resulted from backward slippage of the transcript, induced by the repeat of the trinucle- otide CAC close to the start point (CACCA+,C). The dramatic stability of the CACC complex resembled the properties of true elongation complexes. Thus it was argued that the slippage in fact placed the tetranucleotide into the TBS.

In this work we report another case of slippage of short nascent transcript accompanied by stabilization of the ternary complex that occurs at the phage A P, promoter. Unlike the situation at the rrnB P1 promoter, slippage in the case of A P, does not depend on the repetition of the temdate sequence complementary to the entire short nascent RNA. The ternary complex with slipped RNA has shorter RNA-DNA matching than the complex prior to slippage, but nevertheless it is more stable. The result suggests that TBS is oriented in the ternary complex in such a way that only the two 3’-proximal nucleotides of RNA are base paired to the template strand of DNA.

MATERIALS AND METHODS RNA Polymerase and DNA Templates-RNA polymerase

with histidine-tagged p’ subunit was purified from E. coli R120 strain harboring plasmid pMKa2Ol according to (18). The 280- base pair EcoRI-XhoI DNA fragment (end points, -163 and +123) carrying the phage A P, promoter was excised from the pDN13 plasmid (191, kindly provided by H. Bujard.

In Vitro Dunscription Reactions and Product Analysis- Reactions on the promoter fragment using the immobilized RNA polymerase were performed as described (18). The nucleotide composition of the samples is indicated in the figure legends. The nucleotide concentrations were: ApU, 0.5 mM; nonradioac- tive NTPs, 10 PM, unless indicated otherwise; [(Y-~~PIATP, [CY-~~PICTP, or [a-32P1UTP, 0.1 p~ (3000 Cilmmol). Reactions were incubated for 5 min at room temperature with appropriate nucleotides. To obtain stable ternary complexes, immobilized transcription reactions were washed as described (18).

Electrophoretic analysis of the samples was performed using 12 or 23% gel slabs with acrylamide and N,N’-methylene bisacrylamide proportions indicated in the figure legends. The identity of RNA products was established using differential 32P-labeling, chain-terminating substrates, and synthetic oligo- nucleotide standards as described (17).

RESULTS Stable Ternary Complexes Are Formed on the PL Promoter

During Dunscription Znitiation-The initial sequence of the phage A P, promoter is

1 5 16 C A C A T C A G C A G G A C G C A C U

PPPA APU

CPAPUPC SEQUENCE I

31701

31702 Dunscription Initiation at PL Promoter

FIG. 1. RNA products made on the A. phage A P, promoter during transcription initiation. A, matrix-immobilized RNApolymerase was used to initiate RNA synthesis with limited set of substrates on the P, promoter using ATP as initiator. Whole reactions were applied on the gel in lunes 3-7; only 5% of the reaction volume was loaded in lanes 1-2 to avoid overexposure of the autoradiograph. The samples in lanes 4-7 were washed with large excess of transcription buffer and, where indicated, were ex-

u u

tended by the addition of specified set of G unlabeled NTPs. B , Stable complexes formed with ATP, ApU, and CpApUpC as initiators. The autoradiographs show 23%

tion products. The top part of the gels (not polyacrylamide gel separation of the reac-

The radioactive nucleotides used to label shown) contained no radioactive bands.

chase: B, P.ll.1

1 2 3 4 5 6 7 Initiator: CPAPUPC APU ATP - - A C, A u;c Preincubation:

Chase: : U :,- U : : -

U : -

PPPWPW- b

the reaction products are indicated by boldface.

1 2 3 4 5 6 7 0 9 1 0 1 1

Transcription from this promoter can be initiated with ATP or ApU or tetranucleotide CpApUpC as shown on the scheme. During standard multiple-round abortive initiation reaction in the presence of initiating ATP and [O~-~~PIUTP, the dinucleotide pppApU (boldface indicates the radioactive phosphate) is the only product formed (Fig. IA, lane 1). The abortive initiation reaction in the presence of ATP, UTP, and [(Y-~~PICTP (the "-G" reaction) and the transcription inhibitor rifampicin leads to the formation of the trimer pppApUpC (lane 2). In agreement with the published observations, rifampicin blocks formation of products longer than the trimer (1, 2). The -G reaction with [O~-~*P]UTP in the absence of rifampicin results in the formation of the pppApU, pppApUpC and tetranucleotide pppApUpCpA, which one can expect from the sequence. However, a major additional band of a lower mobility is also formed (lane 3).

Throughout this work we used RNA polymerase tagged with 6 histidines fused to the C terminus of the p' subunit. As is shown elsewhere (18), such RNA polymerase can be immobilized on Ni2+ nitrilotriacetic acid agarose beads without loss of activity, allowing us to perform transcription reaction in solid phase. This system permits easy determination of reaction products that are associated with the transcription complex by washing the beads with an excess of buffer and intermittent brief centrifugation. In Fig. lA, lane 4, the products of the solid phase -G transcription reaction are shown that remain associated with the complex after the washing. The RNA tetramer as well as the slower migrating product were retained on the sorbent, while the dimer and the trimer were completely lost. Surprisingly, the tetramer in the transcription complex could not be extended with GTP, the only nucleotide that was missing in the initial reaction (not shown). Instead, the addition of UTP (lane 5); UTP and CTP (lane 6); and UTP, CTP, and ATP (lane 7) quantitatively extended the tetramer and some of the slower migrating product. The addition of UTP, CTP, and ATP to the stable tetramer resulted in the product with the same mobility as that of the slower migrating band. Thus, the slower migrating band is the heptamer pppApUpCpApUpCpA.

Stable ternary complexes were also formed when ApU or CpApUpC were used to initiate transcription in the presence of CTP and ATP, or ATP only, respectively (Fig. U?, lanes 4 and 9). Again, these complexes could not be extended with GTP, but instead they were quantitatively chased in the presence of UTP (lanes 5 and 10); UTP and CTP (lanes 6 and 11 ); UTP, CTP, and ATP (lane 7), and finally UTP, CTP, ATP, and GTP (lane 8). In the latter case, to prevent the escape into elongation, the Ap-

UpCpA complex was first extended with UTP and CTP, washed, and then extended with ATP and GTP.

Stable Ternary Complexes Are Elongation Proficient-When UTP is omitted from the ApU-initiated reaction on the PL promoter, the RNA polymerases stalls at +15 before the position of the first UMP encountered in the sequence (see Ref 20, Fig. 2A, lane 4). We wished to compare transcripts resulting from the extension of the short RNA in the stable ternary promoter complexes described above with those initiated in the "conventional" +1 mode. For this purpose the stable ApUpCpA complex (lane 1) was extended with UTP (lane 21, washed, and incubated at the "-U" conditions. The resulting complex (lane 3) contained an RNA product with the mobility expected of the 18-mer.

In the experiment of Fig. 2B, the ability of the short RNA in the stable promoter complexes to be "chased" into the runoff product was investigated. In this experiment, ApU-initiated "-G, U" and "-U" complexes, containing the 4-meric and the lbmeric transcripts, respectively, were bead-purified and then labeled with [(Y-~~PIUTP (lanes 1 and 2). (In lane 2, substantial amounts of 25-meric complexes are present, due to readthrough through the U position at +16 and subsequent stalling before U,,,.) The complexes were then allowed to elongate under two different NTP concentrations, and the reaction products were compared on the denaturing gel. The transcripts initiated from the ternary complexes containing short RNA were 3 nucleotides longer than the conventional transcripts. This is most evident from comparison of the paused complexes indicated on the figure. It is noteworthy that the pattern of pauses generated by RNA polymerase molecules started from the two different initial complexes was the same, testifying that the stable complexes containing the short RNA were fully transcriptionally competent.

Dunscription at Undersaturating GTP Concentrations-In the experiment shown in Fig. 3, the effect of varying concentrations of GTP on the appearance of stable ternary complexes containing ApUpCpA was investigated. The solid phase ApU- initiated -U reactions were performed at the indicated concentrations of GTP. After washing, the reaction products associated with transcription complexes were "developed" with radioactive UTP. In the absence of GTP, ApUpCpApU was the only product of the reaction (Fig. 3, lane 1 ). The addition of increasing amounts of GTP resulted in the appearance of the 16-meric product, while the amount of complexes containing the shorter RNA decreased. The apparent K,,, for the synthesis

Dunscription Initiation at PL Promoter 31703

FIG. 2. Two modes of transcription initiation on P, promoter. A, purified transcription complexes were obtained in the ApU-primed reactions a t -U, G and -U reactions (lanes 1 and 4, respectively). The product of the first reaction was extended with UTP in a single step extension reaction (Iane 2 ) , washed, and then extended in the -U reaction. An autoradiograph of the 23% gel (acrylamidei N,N”methylene bisacrylamide, 20:3) is shown. B, purified complexes obtained in the -U, G and -U reactions (lanes 1 and 2 ) were labeled with radioactive UTP and then allowed to resume elongation at the

beled NTPs. An autoradiograph of 12% specified concentration of the four unla-

gel (acrylamide/N,N’-methylene bisacrylamide, 19:l) is shown.

A.

1 2 3 4 5 6 7 8 FIG. 3. Products of the ApU-initiated -U reactions in the pres-

ence of varying concentrations of GTP. The autoradiograph shows RNA synthesized in the immobilized -U reactions at the specified concentrations of GTP. After washing, reaction products stably associated with transcription complexes were developed with radioactive UTP.

of the 16-mer was 1 PM GTP. Properties of the Stable Ternary Promoter Complexes-The

stable ApUpCpA ternary complex was purified, incubated under different salt and temperature conditions, and allowed to elongate RNA in the presence of UTP and CTP. In this assay, appearance of products longer than the initial tetramer indicates that the ternary complex is stable. The results presented in Fig. 4A show that the ApUpCpA-containing ternary complex has a half-life of more than 1 h at room temperature (lane 9); is resistant to rifampicin (lane 3) and low temperature (lanes 10-12); is moderately resistant to salt at room temperature (lanes 6 and 7) ; and is highly salt-resistant at low temperature (lanes 11 and 12). Decreased temperature also significantly increased the half-life of the complex (compare lanes 12 and 9).

In the experiment of Fig. 4B, the stabilities of ternary complexes containing ApUpCpA, ApUpCpApU, and ApUpCpAp- UpC were compared. Surprisingly, extension of stable Ap- UpCpA led to progressively less stable ternary complexes. The pentamer complex had the half-life of 15 min a t room temperature (lane 11 ) and was stable for 30 min on ice (lane 121, while the hexamer complex had the half-life times of 5 min a t room temperature (lane 15) and 30 min on ice (lane 18). The complexes containing longer RNA were even less stable: all of the heptamer ApUpCpApUpCpA and the octomer ApUpCpApUp-

B. R u n 4 -

18-mer 15-mer i]

--- . L

16ma -

u*c I

B. --A~UOC~ADUOCOA “APUPCPApUPC

ApUpCpA ternary complex (lane 1 ) was incubated under indicated con- FIG. 4. Properties of the early ternary complexes. A, purified

ditions and chased into the hexamer by the addition of UTP and CTP. B, purified ternary complexes containing ApUpCpA (lane 1 ), ApUpCpApU (lane 7), and ApUpCpApUpC (lane 13) were challenged with indicated NTPs at the specified time intervals a t two different temperatures. The initial ApUpCpApU complex (lane 7) contains 20% of ApUpCpApUpC complex due to contamination of UTP. A 2 3 8 gel (acrylamide/N,A”- methylene bisacrylamide, 20:3) was used to resolve the reaction products.

CpApG was lost from the complex during the washing proce- dure (not shown).

GreA and GreB Cleavage of Short RNA in the Complexes-In the experiment shown on Fig. 5 purified ternary complexes containing short RNAs were treated with the transcript cleavage factors GreA and GreB. As can be seen, the tetramer Ap- UpCpA was resistant to the factor-induced cleavage (lanes 3 and 5 ) and could be elongated into heptamer in the presence of the factors (lanes 4 and 6). Both GreA and GreB induced cleavage of the pentameric ApUpCpApU and the hexameric ApUpC- pApUpC in the ternary complex to the tetrameric ApUpCpA (lanes 9 and 11 and lanes 15 and 17), which remained associated with transcription complex and could be elongated into heptamer (lanes 10 and 12 and lanes 16 and 18).

31704 Dunscription Initiation at PL Promoter

1 2 3 4 5 6 7 8 9 101112131415161718

FIG. 5. Effect of the Gre factors on the early ternary complexes. Purified ternary complexes containing ApUpCpA (lane 1 1, ApUpCpApU (Iane 7) and ApUpCpApUpC (lane 13) were treated with purified GreA or GreB and then chased by addition of ATP, UTP, and CTP.

DISCUSSION

Noncomplementary Slippage a t the A PL Promoter-While studying early steps of transcription on the phage A P, promoter we observed an unusual stable ternary complex containing the tetrameric transcript ApUpCpA. The short RNA could be extended by successive addition of UTP, CTP, ATP, and GTP. The string UCAG is found only once in the P, promoter-containing DNA fragment used here, namely U+,C+,A+,G+,. There- fore, the results presented in Fig. 1 can be only explained by the model in which the tetrameric RNA is synthesized in the conventional “+1 to +4” mode and then slips backward by 3 nucleotides. In agreement with this model, elongation of stable Ap- UpCpA produces transcripts that are 3 nucleotides longer than the conventional products of the P, promoter. When ATP is used to initiate the -G reaction, the slipped tetramer can be further extended to the heptamer pppApUpCpApUpCpA. Some of the heptameric RNA slips again and could be further elongated by the addition of UTP, CTP, and ATP, but the second slippage occurs less efficiently (Fig. 1A). The two successive transcript slippage events are schematically represented in Fig. 6. By contrast to the previously described instance of slippage at the rrnB P1 promoter (17), slippage at the P, promoter does not require full-length direct repeat in DNA. As the result, the slipped transcript in its new position does not match the topologically corresponding stretch of DNA except for the 3‘- terminal dinucleotide. The absence of stringent sequence con- straint suggests that noncomplementary slippage may be quite general during in vitro initiation of transcription. While the significance of this phenomenon in vivo is unclear, the possi- bility of slippage should be kept in mind while interpret- ing results of in vitro studies of promoter functioning and activation.

Slipped Danscript as a Probe of TBS Topology-Both a t A P, and at the rrnB P1, the slipped ternary complex displays remarkable stability usually characteristic of the elongation stage. As with rrnB P1, the stabilized complexes a t A P, contain u facto? confirming the earlier conclusion that u loss is not required for RNA stabilization. We argue that the stabilization is caused by the entry of the slipped transcript into TBS, the site that is normally involved in stable holding of RNA during processive elongation. Thus the stabilizing tetranucleotide can be viewed as a probe for studying the topology of TBS within the ternary complex. The notion of TBS was first introduced as a functional concept derived from the phenomenon of internal transcript cleavage that removes the 3’-proximal segment of RNA, whereas the 5’ segment remains in the complex (21). TBS was defined as a locality that holds the 5’ segment in the complex. Our present results with GreA- and GreB-induced cleavage in A P, complexes along with similar observations on

* K. Severinov and A. Goldfarb, unpublished observations.

+ AUCA

-10 hexamer * * * * CTATGACTCGTGTAGTCG

PL promoter (template strand) Add U,C,A

t’

CTATGACTCGTGTAGTCG

+ %A’uCA

CTATGACTCGTGTAGTCG * * * * *

AUC Add U,C,A

* t

CTATGACTCGTGTAGTCG

AUC %cA!ucA * *’* * *

CTATGACTCGTGTAGTCG

TBS LBS ”

FIG. 6. ‘lkanscript slippage model explaining early initiation events at the phageA P , promoter. See the text for a discussion.

rrnB Pl(17) delineate TBS as the element accommodating the RNA tetranucleotide immediately upstream from the site of cleavage. A principal observation of this work is that the slipped tetranucleotide is complementary to DNA only at its two downstream positions. This suggests that the stabilizing tetranucleotide in TBS is centered around the point where the nascent transcript branches away from the template DNA. Thus the RNA branching-out point must be separated by two base pairs from the site of internal transcript cleavage.

Extension of Slipped Danscript into the Loose Product Bind- ing Site-In the current model of transcription mechanism ( 8 - 13), the GreB cleavage site defines the boundary between TBS and the downstream loose product binding site (LBS) (Fig. 6). LBS is presumed to constitute the locality where abortive oli- gonucleotides are made during the abortive initiation stage. Filling of LBS with RNA is believed to involve internal move- ment of flexibly connected catalytic center. In support of this model, we have shown that if the priming NTP of the would-be abortive transcript is secured in RNA polymerase by a covalent cross-link, stable products up to 9 nucleotides in length are synthesized as if the catalytic center moves away from the site of the cross-link (22).

The extension of the stabilized slipped tetramer into LBS is topologically equivalent to the extension of a cross-linked NTP, so that the catalytic center advances along LBS as the transcript grows longer. We report here that such process results in destabilization of the transcript so that the longer transcripts have shorter half-lives in the complex. To explain this observation, we invoke the recent evidence that during elongation filling of LBS with RNA a t specific “inchworming“ sequence sig- nals is accompanied by a buildup of intramolecular strain in the ternary complex (10). The strained complex is then “re- laxed” through a forward leaplike translocation accompanied by threading of the transcript through TBS so that LBS is emptied. In the absence of transcription substrates, strained elongation complexes tend to collapse into nonproductive dead- end configuration in which the active center is believed to have lost the RNA 3’ terminus. By analogy, we suggest that at the A P, promoter, extension of the slipped tetramer creates intramolecular strain. Attempts by the complex to relax in the absence of the substrates lead to the loss of the transcript, an event equivalent to the dead-end formation.

fianscription Initiation at PL Promoter 31705

It should be emphasized that TBS, delineated by these ex- periments as the area accommodating the slipped tetranucleotide, may not be the only element holding RNA in the ternary complex during processive elongation. Transcript threading through TBS may involve temporary loosening of the protein grip on RNA. To prevent the loss of the transcript, another RNA holding site further upstream can be envisaged. Together, TBS and the upstream site may provide the mechanism reconciling the phenomenal stability of the elongating complex with the need of rapid extrusion of continuously growing RNA chain.

Acknowledgments-We thank S. B O N ~ ~ O V for providing purified GreA and GreB factors, Bob Landick for communicating his data prior to publication, and Hermann Bujard for providing the P, promoter- containing plasmid.

REFERENCES 1. Johnston, D. E., and McClure, W. R. (1976) in RNA Polymerase (Losick, R. R.,

and Chamberlin, M. J., eds) pp. 101-126, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY

2. Carpousis, A. J., and Gralla, J. D. (1980) Biochemistry 19, 32453253

4. Hansen, U. M., and McClure, W. R. (1980) J. Bid. Chem. 266,9 564-9570 3. Grachev, M. A., and Zaychikov, E. F. (1980) FEBS Lett. 116,23-26

5. 6. 7. 8. 9.

11. 10.

12. 13.

14. 15. 16. 17.

18.

19. 20.

21.

22.

Straney, S. B., and Crothers, D. M. (1987) Biochemistry 26,5063-5070 Metzger, W., Schickor, P., and Heumann, H. (1989) EMBO J. 8,2745-2754 Krummel, B., and Chamberlin, M. J. (1989) Biochemistry 28,7829-7842 Borukhov, S., Sagitov, V., and Goldfarb, A. (1993) Cell 72,459-466 Chamberlin, M. J. (1994) Haruey Lect., Series 88, pp. 1-21, Wliey-Liss, New

Altmann, C. R., Solow-Cordero, D. E., and Chamberlin, M. J. (1994)Proc. Natl. Nudler, E., Goldfarb, A,, and Kashlev, M. (1994) Science 265,793-796

Das, A. (1993) Annu. Reu. Biochem. 62,893-930 Chan, C. L., and Landick, R. (1994) in Dunscription: Mechanisms and Regu-

lation, (Conaway, R. C., and Conaway, J. W., eds) pp. 297-321, Raven Press, Ltd., New York

York

Acad. Sci. U. S. A. 91,37843788

Chamberlin, M., and Berg, P. (1964) J. Mol. Bid. 8, 708-726 Parker, R. C. (1986) Biochemistry 26,65934598 Jacques, J.-P., and Susskind, M. M. (1990) Genes & Den 4, 1801-1810 Borukhov, S., Sagitov, V., Josaitis C. A,, Gourse, R. L., and Goldfarb, A. (1993)

Kashlev, M., Martin, E., Polyakov, A,, Severinov, IC, Nikiforov, V., and

Knaus, R., and Bujard, H. (1988) EMBO J. 7,2919-2923 Kashlev, M., Nudler, E., White, T., Kutter, E., and Goldfarb, A. (1993) Cell 75,

Surrat, C. K., Milan, S. C., and Chamberlin, M. J. (1991) Proc. Natl. Acad. Sci.

Mustaev, A., Kashlev, M., Zaychikov, E., Grachev, M., Goldfarb, A. (1993)

J. Bid. Chem. 268, 23477-23482

Goldfarb, A. (1993) Gene (Amst.) 130,9-14

147-154

U. S. A. 88, 7983-7987

J. Bid. Chem. 268, 19185-19187

topology of the product binding site in rna polymerase revealed

Documents