Eur. J. Biochem. 56, 563 - 569 (1975)
Stabilization of Promoter Complexes with a Single Ribonucleoside Triphosphate Christiane NUSSLEIN and Heinz SCHALLER Max-Planck-Institut fur Virusforschung, Tubingen, und Lehrstuhl fur Mikrobiologie der Universitat Heidelberg (Received February 21 /April 10, 1975)
Under specific binding conditions RNA polymerase forms complexes at several sites of the replicative form DNA of bacteriophage fd. One of these complexes becomes stable to both high salt and low temperature after incubation with GTP. None of the complexes is stabilized by ATP. The stabilization by GTP results from the synthesis of an oligo(G) chain, which is bound in the complex. Size and pyrimidine fingerprints of the DNA segment protected by the enzyme against digestion with DNase are not changed upon initiation of oligo(G) synthesis. This result indicates that binding site and initiation site are identical parts of a promoter region.
RNA polymerase forms very stable complexes at specific sites on the DNA template [2,3]. These sites are parts of protomer regions and contain the initiation point of a strand-selective, specific RNA transcript [4]. We have been studying the RNA polymerase binding sites from the replicative form DNA (RF) of bacteriophage fd. Initial studies indicated that the nucleotide sequences of binding sites derived from different promoter regions on this template are widely different [5]. The sites also differ both in the rate of binding to the polymerase and in the stability of the resulting complexes [1,6]. In order to determine the functionally essential parts of a binding site it is necessary to compare the nucleotide sequences of several such binding sites. The isolation of binding sites as polymerase-bound, DNase-resistant fragments (pDNA) usually leads to a mixture of different binding sites which is difficult to analyse. Because of a higher rate of binding by the polymerase, one fd R F binding site (site I) could be isolated to about 8 5 % purity [ 5 ] . The nucleotide sequence of this binding site has been determined [7]. In an attempt to enrich and isolate other sites we investigated the influence of ribonucleoside triphosphates on the stability of RNA-polymerase . DNA Abbreviations. RF, replicative form DNA; pDNA, polymerasebound, DNase, resistant fragments of the promoter region of DNA. Enzymes. RNA polymerase or nucleoside triphosphate RNA nucleotidyltransferase (EC 2.7.7.6); deoxyribonuclease or pancreatic deoxyribonucleate-oligonucleotidehydrolase (EC 3.1.4.5).
Eur. J. Biochem. 56 (1975)
complexes. It has been reported that the presence of a single ribonucleoside triphosphate stabilizes RNApolymerase . DNA complexes against decay at high ionic strength [8,9]. Since ATP and GTP are most effective in this respect, it has been suggested that the stabilization is due to the formation of ternary “initiation complexes” of polymerase, DNA and the first ribonucleoside triphosphate. In this paper we present evidence that stabilization by purine triphosphates is not a general feature of RNA-polymerase . DNA complexes, but occurs only in special cases due to the synthesis of a homopolymer chain within the complexes. Applications of this stabilization for the isolation of RNA polymerase binding sites will be discussed.
MATERIALS AND METHODS Many of the materials and methods used have been described in the preceding paper [4]. Ribonucleoside triphosphates were purified by ion-exchange chromatography on DEAE-Sephadex A-25. The purification of triphosphate did not significantly alter the results obtained. Calf thymus DNA (Worthington) was dissolved in Tris-EDTA buffer to a concentration of 3 mg/ml and denatured by heating to 95 “C for 5 min. Nitrocellulose filters, 0.15-pm pore size, 2.5-cm diameter, were from Sartorius Membranfilter-Gesellschaft (Gottingen). They were soaked in binding
564
Stabilization of RNA-Polymerase . DNA Complexes
buffer (buffer B) and 40 mM KCl for at least 10 rnin before use. Buffer E is 20 mM Tris . HCl pH 8.0, 0.3 mM dithiothreitol, 1 mM EDTA, 5% glycerol. Buffer B is buffer E 10 mM MgC1,.
+
Filter Binding Assay RNA polymerase and fd RF were mixed at 0 "C in buffer B and 200 pg/ml bovine serum albumin and 120 mM KCl and incubated for 10 rnin at 37 "C. Denatured calf thymus DNA was added to a concentration of 300 pg/ml. After a further 10 rnin at 37 "C the sample was diluted into 0.5 ml buffer B and 40 mM KC1 at 37 "C and filtered through nitrocellulose filters (2.5-mm diameter). The filters were washed with 1 ml of the same buffer, dried and counted in a toluene-based scintillation fluid. To assay complex stabilization, an incubation with ribonucleoside triphosphates as indicated was included after the addition of calf thymus DNA. The mixture was diluted with 0.5 ml buffer E and 200 mM KC1 at 0 "C, kept at 0 "C for 10 min and filtered as described above. Any variations of this procedure are given in the figure legend. Isolation of Protected DNA 7 pmol 32P-labelled fd RF (6.6 x lo6 counts/min) were incubated with RNA polymerase as indicated in 250 pl buffer B and 200 pg/ml bovine serum albumin and 120 mM KC1 for 10 min at 37 "C. Denatured calf thymus DNA (final concentration 300 pg/ml) was added and after 5 min either GTP (final concentration 50 pM), or ATP (to 150 pM) and UTP (to 300 pM) was added. After another 5 rnin at 37 "C the mixture was incubated at 0 "C for 15 min. It was then prewarmed to 37 "C and diluted with an equal volume of 100 pg/ml DNase solution. After 10 rnin at 37 "C the digestion was stopped by the addition of EDTA to a final concentration of 30 pM. The mixture was fractionated on a Sephadex G-75 column (0.8 x 30 cm) equilibrated with 0.01 M Tris . HCl pH 7.5 and 0.1 mM EDTA and 40 mM KC1. The fractions from the void volume were pooled, 0.1 vol. of 1 M ammonium acetate pH 5.0 and 2 vol. of ethanol were added and after at least 2 h at - 20 "C the precipitate was collected by centrifugation in the SS34 rotor of the Sorvall centrifuge (10000 rev./min, 10 min). The precipitate was dissolved in 200 pl 0.1 M Tris base and extracted with an equal volume of 80% phenol. The nucleic acids were precipitated from the aqueous phase with ethanol as described above. The precipitate was dissolved in 20 pl H20. For the isolation of DNA protected in non-stabilized complexes, the bovine serum albumin was omitted from the binding mixture
and the DNase digestion step preceded the incubation at 0 "C. After incubation with DNase, the mixture was adjusted to 30mM EDTA, 200mM KC1 and 300 pg/ml denatured calf thymus DNA and incubated at 0 "C for 10 min. It was diluted with 0.5 ml buffer E and 40 mM KCl and filtered through a nitrocellulose filter. The filter was washed with 1 ml of the same buffer and the DNase-resistant DNA recovered from the combined filtrates as described above. RESULTS Stabilization of RNA-Polymerase .fd-RF Complexes with GTP RNA-polymerase . DNA complexes are very stable at 37 "C and low ionic strength, whereas decay is rapid when the salt concentration is raised or the temperature is lowered [lo, 31. This dissociation of fd-RF . RNA-polymerase complexes under restrictive conditions can be prevented by preincubation with GTP. Table 1 shows the effect that preincubation of RNA-polymerase . fd-RF complexes and ribonucleoside triphosphates has on the stability of the complexes in 200 mM KC1 at 0 "C. Without triphosphates, essentially all complexes dissociate within 10 min, whereas after incubation with GTP they Table 1. Stabilization of RNA-polymerase .fd-RF complexes with ribonucleoside triphosphates 3.3 pmol fd RF (40000 counts 14C/min) and RNA polymerase ( r = 10) were incubated in 400 p1 of buffer B 0.2 mg bovine serum albumin/ml + 120 mM KC1 for 10 min at 37 "C. Denatured calf thymus DNA was added, and after 10 rnin 25.~1 aliquots were taken and added to a 25-pl ribonucleoside triphosphate solution (final concentration of all triphosphates 100 pM). After 10 rnin at 37 "C the samples were diluted with 0.5 ml buffer B containing 200 mM KC1 and after 10min at 0 ° C filtered through nitrocellulose filters. Results were corrected for background obtained in absence of nucleoside triphosphates (171 counts/min). Percentage of stabilized complexes was calculated relative to complexes retained onnitrocellulose after binding and competitionat 37 "C (1850counts/ min)
+
Ribonucleoside triphosphate
ATP GTP UTP CTP ATP ATP ATP GTP GTP UTP ATP
+ GTP + UTP + CTP + UTP + CTP + CTP + UTP + CTP + GTP
Complexes retained on filter counts/min
x
49 1379 101 36 1371 413 253 1166 1372 219 1362
3.3 85 6.6 2.2 84 25 15.5 70 84 13 84
Eur. J. Biochem. 56 (1975)
C. Nusslein and H. Schaller
565
5
0
10
RNA polymerase / f d RF ( mol/rnol)
Fig. 1. Stabilization of RNA-polymerase .fd-RF complexes with GTP. RNA-polymerase . fd-RF complexes were formed at different polymerase/fd R F ratios in 75 p1 buffer B + 0.2 mg bovine serum 120 mM KC1 for 30 min at 37 ‘C. The R F concentraalbumin tion was 2.6pmoljml. Denatured calf thymus DNA was added to a concentration of 300 pgjml and after a further 10 min at 37 “C, 25-pI samples were diluted with 0.5 ml buffer B + 40 mM Other samples were incubated KC1 and filtered directly (-0). for further 10 min with 100 pM GTP (0-O), or 100 pM ATP (A-A), or without triphosphates (A---A) and processed as described in the legend to Table 1
+
0
remain stable. ATP, UTP or CTP have essentially no stabilizing effect. Incubation with a mixture of two ribonucleoside triphosphates leads to stable complexes whenever GTP is one of the triphosphates. A small, but significant,effect is also observed with the combination of ATP and UTP. At the salt concentration used for the binding reaction in our experiments (120 mM KCl), four different sites of the replicative form DNA can be stably complexed by RNA polymerase. At two of these sites the RNA chains are initiated with ATP, at the other two with GTP [l]. We have shown that under these conditions (i.e. at 120 mM KC1) the binding of RNA polymerase to the different sites occurs at different rates. At low ratios of polymerase to DNA ( r ) only the “fast” binding site I is bound by polymerase, whereas at higher ratios the “slow” binding sites on the replicative form DNA are also bound [5,1]. Upon addition of ribonucleoside triphosphates, complexes formed at binding site I initiate an RNA chain starting with GTP [4,11]. In order to find out whether GTP stabilises complexes with “fast” or “slow” binding sites, the stabilization of complexes formed at increasing r was measured. Fig. 1 shows that complexes formed at low, as well as at high r, are stabilized to 80- 90 % by incubation with GTP, indicating that at least the complexes of RNA polymerase with binding site I are a target for GTP. The experiment can not measure a possible GTP stabilization at the other three binding sites on the DNA molecule, since complex formation at one site is presumably enough for preventing the molecule from passing through the filter. ATP has little effect on complexes, even at high r, although under these conditions two sites which specify ATP-initiated RNA chains are bound by the polymerase [l]. In order to find out whether complexes other than that at binding site 1 are stabilized with GTP, the total amount of DNA protected by polymerase against DNase digestion was determined. Fig. 2 shows that even at high r values the protected portion does not exceed 0.6% of the replicative form DNA. This corresponds to one site (40 base pairs) per DNA molecule (6000 base pairs) and indicates that only the complex at binding site I is stabilized with GTP. This result is confirmed by the pyrimidine tract analysis of the protected DNA, shown below. It should be noted that the above experiments include a direct measurement of the efficiency of the filter binding assay. As shown in Fig. 1, DNA molecules complexed by RNA polymerase at one site only are retained by nitrocellulose to about 85 %. The actual efficiency of retention may even be higher, since not all complexes at this site may been stabilized by incubation with GTP. A lowered capacity
: 5 10 RNA polymerase/fd RF(mol/rnol)
Fig. 2. Isolation of G TP-stabilized RNA-polymerase . p D N A complexes. RNA-polymerase . fd-RF complexes were formed as described in the legend to Fig.l. The R F concentration was 0.58 pmol (5 x lo5 counts 32P/min) in 75 pl. After incubation with or without triphosphates (-0, A-A) GTP (0-0) 25-pl samples were digested with 25 pl DNase solution (100 pg/ml 0.5 mM CaCl,) for 10 min at 37 “C, and filtered in buffer B through nitrocellulose filters (0-0) or diluted with 0.5 ml 200 mM KCI 30 pg/ml denatured calf thymus DNA, buffer E A-A) kept for 10 min at 0 “C and filtered (0-0,
+
+
+
Eur. J. Biochem. 56 (1975)
566
Stabilization of RNA-Polymerase
DNA Complexes
i ; ( / j / i ( , O / / J ~ J / I . / ~ ~ [ , ~ ~ / , s ~ , - / ~ ~ ( ~DIVA. / ~ ~ ~ / ( , ~DNA / protected i n (A) GTP-stabilized complesca 0'-= 5 ) . (U) RNAFig. 3. ~ ~ / ; , ~ ( ~ / J l , ~ i ~ ~/J(///?UI.\ polymerase . I'd-RF complexes ( I ' = 5). (C) complexes not stabilized with G T P at I' = 10 (not retained by nitrocellulose filters after incubation at 0 C ) . (11) complexes stabilized with ATP + UTP ( r = lo), was isolated as described in Materials and Methods. The fd RF used was labelled i n the plus strand with 3zP. pDNA was depurinised and the resulting pyrimidine tracts separated by two-dimensional thin-layer chromatography as described [5]
(,30-70":,) for GTP stabilization (and for KNA chain initiation) has been observed with aged preparations of RNA polymerase, although the enzyme appeared to be normal with respect to specific DNA binding. Despite the high efficiency of the binding assay, about two enzyme molecules were always required to reach the plateau of GTP-stabilized complexes, i.e. to complex one binding site. This result was obtained with all (four) enzyme preparations used. It could be caused by (a) a partially inactive enzyme, (b) an overestimation of the protein concentration, or (c) binding of the RNA polymerase in its dimeric form. The latter explanation seems to be less likely, since isolated binding-site . polymerase complexes have been shown to contain the enzyme monomer (U. Jellinghaus, unpublished results). Isolniiori of ProtecttJd D N A ,from Stabilized Complexes
The selective stabilization of binding-site . polymerase complexes offers several possibilities for the
purification of binding sites. For the preparative isolation of the DNA from GTP-stabilized complexes, complexes were formed as described in the legends to Table 1 and Fig. 1 ( r = 5). After incubation with GTP and then at 0 "C the replicative form DNA was digested with DNase and the protected DNA was isolated as described in Methods. The pyrimidine tract analysis (Fig. 3 A) indicates the presence of only the oligopyrimidines characteristic for binding site I [5,7]. The pattern of a pDNA preparation isolated in a parallel experiment ( r = 5 , no incubation with GTP, no incubation at 0 'C) is shown in Fig. 3B. The fingerprint is more complex and corresponds to a mixture of several binding sites [ 5 ] . The comparison of the two fingerprints shows that all complexes formed at these sites dissociated during the incubation at 0 " C , except for the one formed at binding site I . The protocol of the experiment described in Fig. 2 allows the separation of binding site I from the other binding sites and thus the isolation of sites whose complexes are not stabilized with GTP. For this Cur. J . Biocheni. .i6 (1975)
C. Niisslein and H. Schaller
purpose the DNase digestion was done before the incubation at 0 "C. The DNase was inactivated by addition of EDTA and non-stabilized complexes were dissociated by incubation at 0 "C in 200 mM KCl. The resulting mixture of pDNA and pDNA . RNApolymerase complexes was fractionated by nitrocellulose filtration. The complexes remained on the filter while the pDNA appeared in the filtrate. The pyrimidine pattern is shown in Fig. 3C. As expected, the spots characteristic for binding site I (namely TC3, T3C2 and T5) [5,7] are missing. Binding site I can be eluted from the filter. In an attempt to extend the method of selective stabilization of binding-site . polymerase complexes for sites other than binding site I, we isolated complexes which had been stabilized by the combination of ATP and UTP. The pattern of this DNA (Fig. 3 D) is similar to the one shown in Fig. 3C, in that again the spots characteristic for binding site I [5,7] are missing. The complexity of the pattern, however, indicates the presence of more than one site in this preparation, suggesting that two ATP-starting complexes have been stabilized. Since the concentration of ATP and UTP required for an efficient stabilization is rather high (100 pM ATP and 200 pM UTP are required to stabilize 90% of the complexes in 10 min at 37 "C) perhaps DNA fragments from RNA synthesizing complexes might also be enriched in this preparation, resulting in an unspecific background. Whether poly(A, U) is synthesized in complexes stabilized by ATP and UTP, and if so how long such chains must be to result in stabilization, is still being investigated.
Mechanism of Stabilization with GTP The finding that only one out of four RNApolymerase . binding-site complexes can be stabilized by incubation with a single ribonucleoside triphosphate makes it highly improbable that the stabilization is due to the formation of "initiation complexes" with an altered conformation, as has been proposed by Anthony et al. [S]. The experiments to be described in this section suggest instead that the stability of the complexes is caused by the synthesis of a short oligo(G) chain within the RNA-polymerase . binding-site complex. GTP-stabilized complexes behave similarly to RNA-synthesizing complexes, which are also stable at high ionic strength and low temperature (see Table 1, last line, and [12]). The stabilization is prevented under conditions which also prevent synthesis of RNA and oligo(G) but which do not impair the binding reaction, e.g. preincubation with rifampicin or incubation without Mg2+ ions [3]. Complexes Eur. J. Biochem. 56 (1975)
567
Time ( r n i n ) Fig. 4. Incorj~oratioiz of / ' H J C ' T P iriro RNA-polymerase .fd-RF complexes. 2.5 pmol 32P-labelledfd RF (7000 counts min-' pmol- ') were incubated with 4.2 pmol RNA polymerase in 100 pI binding buffer containing 200 pg bovine serum albumin/ml and 120 mM KCI for 10 min at 37 "C. 20 pI of a [3H]GTP solution in the same buffer (60 pM GTP, 7600 counts min-' pmol -') was added and further incubated at 37 'C. 6-pl samples were taken, diluted in 0.5 ml buffer E containing 200 mM KC1, 10 pM GTP, 3 pg/ml calf thymus D N A at 0 "C. After 10 min at 0 "C the samples were filtered through nitrocellulose filters. A background of 170 counts 3H/min and 70 counts 32P/minwas substracted. (0-0) [32P]RF; (O--O) oligo([3H]G)
once stabilized, however, are insensitive to rifampicin and do not dissociate even if the Mg2+ ions are complexed by addition of excessive EDTA (data not shown). This suggests that it is the product of the reaction which renders the complexes stable. In order to detect a reaction product, labelled GTP was added to preformed RNA-polymerase . fdR F complexes. The kinetics of the stabilization in the presence of 10 pM t3H]GTP is shown in Fig. 4. In the initial period GTP label corresponding to approximately 4 GMP residues per stabilized complex is retained by the filter. After 10 min, when 90% of the complexes are stabilized, additional GMP residues are incorporated with reduced rate. One might argue that the stabilization by GTP reflects the synthesis of a normally initiated RNA product, caused by a contamination of GTP with other ribonucleoside triphosphates, rather than by the synthesis of oligo(G). Since, however, the GTP concentration required for stabilization is very low and purified GTP has the same effect, this possibility is excluded. Thus we conclude that the presence of GTP in the complex of RNA polymerase and binding site I allows an oligo(G) chain to be synthesized. This oligo(G) chain remains bound in the complex and alters its conformation, so that it no longer dissociates
568
Stabilization of RNA-Polymerase . DNA Complexes
under conditions restrictive for RNA-polymerase . DNA complexes. DISCUSSION The experiments described in the present paper show that RNA-polymerase . DNA complexes can be stabilized in special cases by preincubation with a single ribonucleoside triphosphate. It is shown that the stabilization is due to the formation of an RNA product which remains bound in a ternary complex with the enzyme and the DNA template. The same conclusion has also been drawn by McConnel and Bonner [13]who showed that apurification of the ribonucleoside triphosphates greatly reduces their ability to stabilize T7-DNA . RNA-polymerase complexes, and by So and Downey [14] who
start point (i.e. the CMP residue on the template strand coding for the starting GTP of the RNA chain) lies at the end of a C-C-C sequence approximately in the center of the binding site. This C cluster most likely acts as template for oligo(G) chain synthesis, elongation of the chain being mediated by a slippage of the product on the template [15]. This model postulates that phosphodiester bond formation can also occur next to the start point (i.e. between position - 1 and 1 instead of 1 and 2, see scheme below). Such bond formation is also indicated by the occurrence of RNA chains starting with ppp(Gp),UpA. (n = 1- 4) under apparently normal conditions for RNA synthesis (see the preceding paper [4]). Similarly RNA chain initiation next to the start point can be forced by dinucleotides complementary to the template strand, but not covering the start point [16].
+
+
+
PPPG PPPU PPPG 1 PPPG 1 Minus strand 3‘ . . . G-T-C-C-C-A-T-T . . . . . . G-T-C-C-C-A-T-T . . . 5’ -4 -3-2 -1 1 2 3 4 -4 -3-2 -1 1 2 3 4 ~~~
Initiation found that on a poly[d(A-T)] template the presence of ATP only was not sufficient to stabilize the complexes. Thus initiation complexes (defined as ternary complexes of RNA polymerase, DNA and initiating nucleoside triphosphate) do not possess a special resistance against high ionic strength, as has been speculated previously [S]. The presence of an RNA chain in the stabilized complex could alter the conformation of the DNA template by annealing to a short region of the template strand and thus stabilizing the denatured part of the DNA segment covered by RNA polymerase. This alteration of the template conformation may explain the completely altered behaviour of RNA-polymerase . binding-site complexes with respect to ionic strength and temperature. During the oligo(G) chain synthesis the polymerase apparently does not alter its position on the template since the pyrimidine tract analysis, as well as the chain length determination [3], reveal no difference between the DNA segments complexed before and after stabilization with GTP. This indicates that the oligo(G) chain synthesis takes place very close to the start point of protomer region I. The question now arises as to why, in this complex, a homopolymer synthesis takes place, while in several other complexes such a synthesis apparently does not occur. From the sequence analysis of binding site I [7] and the initiation sequence of the RNA directed by protomer region I [4], we know that the
Oligo(G) synthesis According to this explanation, stabilization by a single ribonucleoside triphosphate depends on the clustering of identical nucleotide residues at the start point. In fd replicative form DNA only complexes at one of the four sites which are stably bound by RNA polymerase at high ionic strength (120 mM KCI) can be stabilized with GTP. In addition a site which is stably bound only at low ionic strength is also stabilized under the appropriate conditions [l].No complex is stabilized with ATP or any other ribonucleoside triphosphate. Unpublished experiments indicate that complexes of RNA polymerase with diX174 replicative form DNA can be stabilized with ATP or CTP, but not with GTP. This agrees with the pppA-A-A . . . and pppC-C . . . initiation sequences observed recently in @XmRNA in vitvo [17]. Selective GTP stabilization has been also reported with a promoter region from T3 DNA [181. Thus the stabilization with single ribonucleoside triphosphates appears not to be restricted to oligo(G) synthesis and to the fd system. Although not generally applicable it provides a valuable tool for the isolation of single RNA polymerase binding sites in a pure form and high yield. Finally we may ask whether oligo(G) synthesis also occurs under physiological conditions, i.e. in the presence of the other nucleoside triphosphates. GTP stabilization, and thus also oligo(G) synthesis, is significantly reduced by the presence of UTP (Table l), probably by the incorporation of an UMP residue which prevents the “backwards slippage” of the Eur. J . Blochem. 56 (1975)
C. Nusslein and H. Schaller
newly synthesized oligonucleotide on the template. Nevertheless, even in the presence of all four nucleoside triphosphates about half of the RNA chains initiated at promotor region I were found to contain two to four G residues at their 5' terminus [4]. This result indicates that oligo(G) synthesis proceeds with a rate comparable to that of the pppG-U . . . initiation. It also shows that the oligo(G) chain produced can be utilized effectively as a primer for RNA chain growth. This interpretation is also supported by the finding that RNA chains can be initiated at promoter region I with high preference by addition of CTP, UTP, and ATP to preformed, GTP-stabilized RNApolymerase . fd-RF complexes (unpublished results).
REFERENCES 1. Seeburg, P. H. & Schaller, H. (1975) J . Mol. Biol. 92, 261277. 2. Hinkle, D. & Chamberlin, M. (1970) Cold Spring Harbor Symp. Quant. Biol. 35, 65 - 72. 3. Niisslein, C. (1973) Ph. D. Thesis, Tubingen.
569 4. Heyden, B., Nusslein, C. & Schaller, H. (1975) Eur. J . Biochern. 55, 147-155. 5. Heyden, B., Nusslein, C. & Schaller, H. (1972) Nut. New B i d . 240,9- 12. 6. Seeburg, P. H. (1975) Ph. D . Thesis, Tubingen. 7. Schaller, H., Gray, C. P. & Herrmann, K. (1975) Proc. Nut1 Acad. Sci. U.S.A. 72, 737-741. 8. Anthony, D. D., Zeszotek, E. & Goldwaith, D . A. (1966) Proc. NatlAcad. Sci. U.S.A. 56, 1026-1033. 9. Stead, N. W. & Jones, 0. W. (1967) J . Mol. Biol. 26, 131- 135. 10. Hinkle, D. & Chamberlin, M. (1972) J . Mol. Biol. 70, 157185. 11. Heyden, B. (1974) Ph. D. Thesis, Tubingen. 12. Naito, S. & Ishihama, A. (1973) Biochem. Biophys. Res. Commun. 51,323 - 330. 13. Mc Connell, D . J. & Bonner, J. (1973) Eur, J . Biochem. 38, 111 -121. 14. So, A. G. & Downey, K. M. (1970) Biochemistry, 9, 47884793. 15. Chamberlin, M. & Berg, P. (1962) Proc. Nut1 Acad. Sci. U.S.A. 48,81-94. 16. Minkley, E. G. & Pribnow, D . (1973) J . Mol. Bid. 77, 255277. 17. Smith, L. H., Grohmann, K. & Sinsheimer, R. L. (1974) Nucleic Acids Research, 1, 1521- 1529. 18. Takeya, T. & Takanami, M. (1974) Biochemistry, 13, 53885394.
C. Nusslein, Biozentrum der Universitat Basel, Klingelbergstrasse 70, CH-4056 Basel, Switzerland
H. Schaller *, Mikrobiologie der Ruprecht-Karl-Universitat Heidelberg, D-6900 Heidelberg 1, Im Neuenheimer Feld 280, Federal Republic of Germany
* To whom correspondance should be sent
Eur. J. Biochem. 56 (1975)
Stabilization of Promoter Complexes with a Single Ribonucleoside Triphosphate Christiane NUSSLEIN and Heinz SCHALLER Max-Planck-Institut fur Virusforschung, Tubingen, und Lehrstuhl fur Mikrobiologie der Universitat Heidelberg (Received February 21 /April 10, 1975)
Under specific binding conditions RNA polymerase forms complexes at several sites of the replicative form DNA of bacteriophage fd. One of these complexes becomes stable to both high salt and low temperature after incubation with GTP. None of the complexes is stabilized by ATP. The stabilization by GTP results from the synthesis of an oligo(G) chain, which is bound in the complex. Size and pyrimidine fingerprints of the DNA segment protected by the enzyme against digestion with DNase are not changed upon initiation of oligo(G) synthesis. This result indicates that binding site and initiation site are identical parts of a promoter region.
RNA polymerase forms very stable complexes at specific sites on the DNA template [2,3]. These sites are parts of protomer regions and contain the initiation point of a strand-selective, specific RNA transcript [4]. We have been studying the RNA polymerase binding sites from the replicative form DNA (RF) of bacteriophage fd. Initial studies indicated that the nucleotide sequences of binding sites derived from different promoter regions on this template are widely different [5]. The sites also differ both in the rate of binding to the polymerase and in the stability of the resulting complexes [1,6]. In order to determine the functionally essential parts of a binding site it is necessary to compare the nucleotide sequences of several such binding sites. The isolation of binding sites as polymerase-bound, DNase-resistant fragments (pDNA) usually leads to a mixture of different binding sites which is difficult to analyse. Because of a higher rate of binding by the polymerase, one fd R F binding site (site I) could be isolated to about 8 5 % purity [ 5 ] . The nucleotide sequence of this binding site has been determined [7]. In an attempt to enrich and isolate other sites we investigated the influence of ribonucleoside triphosphates on the stability of RNA-polymerase . DNA Abbreviations. RF, replicative form DNA; pDNA, polymerasebound, DNase, resistant fragments of the promoter region of DNA. Enzymes. RNA polymerase or nucleoside triphosphate RNA nucleotidyltransferase (EC 2.7.7.6); deoxyribonuclease or pancreatic deoxyribonucleate-oligonucleotidehydrolase (EC 3.1.4.5).
Eur. J. Biochem. 56 (1975)
complexes. It has been reported that the presence of a single ribonucleoside triphosphate stabilizes RNApolymerase . DNA complexes against decay at high ionic strength [8,9]. Since ATP and GTP are most effective in this respect, it has been suggested that the stabilization is due to the formation of ternary “initiation complexes” of polymerase, DNA and the first ribonucleoside triphosphate. In this paper we present evidence that stabilization by purine triphosphates is not a general feature of RNA-polymerase . DNA complexes, but occurs only in special cases due to the synthesis of a homopolymer chain within the complexes. Applications of this stabilization for the isolation of RNA polymerase binding sites will be discussed.
MATERIALS AND METHODS Many of the materials and methods used have been described in the preceding paper [4]. Ribonucleoside triphosphates were purified by ion-exchange chromatography on DEAE-Sephadex A-25. The purification of triphosphate did not significantly alter the results obtained. Calf thymus DNA (Worthington) was dissolved in Tris-EDTA buffer to a concentration of 3 mg/ml and denatured by heating to 95 “C for 5 min. Nitrocellulose filters, 0.15-pm pore size, 2.5-cm diameter, were from Sartorius Membranfilter-Gesellschaft (Gottingen). They were soaked in binding
564
Stabilization of RNA-Polymerase . DNA Complexes
buffer (buffer B) and 40 mM KCl for at least 10 rnin before use. Buffer E is 20 mM Tris . HCl pH 8.0, 0.3 mM dithiothreitol, 1 mM EDTA, 5% glycerol. Buffer B is buffer E 10 mM MgC1,.
+
Filter Binding Assay RNA polymerase and fd RF were mixed at 0 "C in buffer B and 200 pg/ml bovine serum albumin and 120 mM KCl and incubated for 10 rnin at 37 "C. Denatured calf thymus DNA was added to a concentration of 300 pg/ml. After a further 10 rnin at 37 "C the sample was diluted into 0.5 ml buffer B and 40 mM KC1 at 37 "C and filtered through nitrocellulose filters (2.5-mm diameter). The filters were washed with 1 ml of the same buffer, dried and counted in a toluene-based scintillation fluid. To assay complex stabilization, an incubation with ribonucleoside triphosphates as indicated was included after the addition of calf thymus DNA. The mixture was diluted with 0.5 ml buffer E and 200 mM KC1 at 0 "C, kept at 0 "C for 10 min and filtered as described above. Any variations of this procedure are given in the figure legend. Isolation of Protected DNA 7 pmol 32P-labelled fd RF (6.6 x lo6 counts/min) were incubated with RNA polymerase as indicated in 250 pl buffer B and 200 pg/ml bovine serum albumin and 120 mM KC1 for 10 min at 37 "C. Denatured calf thymus DNA (final concentration 300 pg/ml) was added and after 5 min either GTP (final concentration 50 pM), or ATP (to 150 pM) and UTP (to 300 pM) was added. After another 5 rnin at 37 "C the mixture was incubated at 0 "C for 15 min. It was then prewarmed to 37 "C and diluted with an equal volume of 100 pg/ml DNase solution. After 10 rnin at 37 "C the digestion was stopped by the addition of EDTA to a final concentration of 30 pM. The mixture was fractionated on a Sephadex G-75 column (0.8 x 30 cm) equilibrated with 0.01 M Tris . HCl pH 7.5 and 0.1 mM EDTA and 40 mM KC1. The fractions from the void volume were pooled, 0.1 vol. of 1 M ammonium acetate pH 5.0 and 2 vol. of ethanol were added and after at least 2 h at - 20 "C the precipitate was collected by centrifugation in the SS34 rotor of the Sorvall centrifuge (10000 rev./min, 10 min). The precipitate was dissolved in 200 pl 0.1 M Tris base and extracted with an equal volume of 80% phenol. The nucleic acids were precipitated from the aqueous phase with ethanol as described above. The precipitate was dissolved in 20 pl H20. For the isolation of DNA protected in non-stabilized complexes, the bovine serum albumin was omitted from the binding mixture
and the DNase digestion step preceded the incubation at 0 "C. After incubation with DNase, the mixture was adjusted to 30mM EDTA, 200mM KC1 and 300 pg/ml denatured calf thymus DNA and incubated at 0 "C for 10 min. It was diluted with 0.5 ml buffer E and 40 mM KCl and filtered through a nitrocellulose filter. The filter was washed with 1 ml of the same buffer and the DNase-resistant DNA recovered from the combined filtrates as described above. RESULTS Stabilization of RNA-Polymerase .fd-RF Complexes with GTP RNA-polymerase . DNA complexes are very stable at 37 "C and low ionic strength, whereas decay is rapid when the salt concentration is raised or the temperature is lowered [lo, 31. This dissociation of fd-RF . RNA-polymerase complexes under restrictive conditions can be prevented by preincubation with GTP. Table 1 shows the effect that preincubation of RNA-polymerase . fd-RF complexes and ribonucleoside triphosphates has on the stability of the complexes in 200 mM KC1 at 0 "C. Without triphosphates, essentially all complexes dissociate within 10 min, whereas after incubation with GTP they Table 1. Stabilization of RNA-polymerase .fd-RF complexes with ribonucleoside triphosphates 3.3 pmol fd RF (40000 counts 14C/min) and RNA polymerase ( r = 10) were incubated in 400 p1 of buffer B 0.2 mg bovine serum albumin/ml + 120 mM KC1 for 10 min at 37 "C. Denatured calf thymus DNA was added, and after 10 rnin 25.~1 aliquots were taken and added to a 25-pl ribonucleoside triphosphate solution (final concentration of all triphosphates 100 pM). After 10 rnin at 37 "C the samples were diluted with 0.5 ml buffer B containing 200 mM KC1 and after 10min at 0 ° C filtered through nitrocellulose filters. Results were corrected for background obtained in absence of nucleoside triphosphates (171 counts/min). Percentage of stabilized complexes was calculated relative to complexes retained onnitrocellulose after binding and competitionat 37 "C (1850counts/ min)
+
Ribonucleoside triphosphate
ATP GTP UTP CTP ATP ATP ATP GTP GTP UTP ATP
+ GTP + UTP + CTP + UTP + CTP + CTP + UTP + CTP + GTP
Complexes retained on filter counts/min
x
49 1379 101 36 1371 413 253 1166 1372 219 1362
3.3 85 6.6 2.2 84 25 15.5 70 84 13 84
Eur. J. Biochem. 56 (1975)
C. Nusslein and H. Schaller
565
5
0
10
RNA polymerase / f d RF ( mol/rnol)
Fig. 1. Stabilization of RNA-polymerase .fd-RF complexes with GTP. RNA-polymerase . fd-RF complexes were formed at different polymerase/fd R F ratios in 75 p1 buffer B + 0.2 mg bovine serum 120 mM KC1 for 30 min at 37 ‘C. The R F concentraalbumin tion was 2.6pmoljml. Denatured calf thymus DNA was added to a concentration of 300 pgjml and after a further 10 min at 37 “C, 25-pI samples were diluted with 0.5 ml buffer B + 40 mM Other samples were incubated KC1 and filtered directly (-0). for further 10 min with 100 pM GTP (0-O), or 100 pM ATP (A-A), or without triphosphates (A---A) and processed as described in the legend to Table 1
+
0
remain stable. ATP, UTP or CTP have essentially no stabilizing effect. Incubation with a mixture of two ribonucleoside triphosphates leads to stable complexes whenever GTP is one of the triphosphates. A small, but significant,effect is also observed with the combination of ATP and UTP. At the salt concentration used for the binding reaction in our experiments (120 mM KCl), four different sites of the replicative form DNA can be stably complexed by RNA polymerase. At two of these sites the RNA chains are initiated with ATP, at the other two with GTP [l]. We have shown that under these conditions (i.e. at 120 mM KC1) the binding of RNA polymerase to the different sites occurs at different rates. At low ratios of polymerase to DNA ( r ) only the “fast” binding site I is bound by polymerase, whereas at higher ratios the “slow” binding sites on the replicative form DNA are also bound [5,1]. Upon addition of ribonucleoside triphosphates, complexes formed at binding site I initiate an RNA chain starting with GTP [4,11]. In order to find out whether GTP stabilises complexes with “fast” or “slow” binding sites, the stabilization of complexes formed at increasing r was measured. Fig. 1 shows that complexes formed at low, as well as at high r, are stabilized to 80- 90 % by incubation with GTP, indicating that at least the complexes of RNA polymerase with binding site I are a target for GTP. The experiment can not measure a possible GTP stabilization at the other three binding sites on the DNA molecule, since complex formation at one site is presumably enough for preventing the molecule from passing through the filter. ATP has little effect on complexes, even at high r, although under these conditions two sites which specify ATP-initiated RNA chains are bound by the polymerase [l]. In order to find out whether complexes other than that at binding site 1 are stabilized with GTP, the total amount of DNA protected by polymerase against DNase digestion was determined. Fig. 2 shows that even at high r values the protected portion does not exceed 0.6% of the replicative form DNA. This corresponds to one site (40 base pairs) per DNA molecule (6000 base pairs) and indicates that only the complex at binding site I is stabilized with GTP. This result is confirmed by the pyrimidine tract analysis of the protected DNA, shown below. It should be noted that the above experiments include a direct measurement of the efficiency of the filter binding assay. As shown in Fig. 1, DNA molecules complexed by RNA polymerase at one site only are retained by nitrocellulose to about 85 %. The actual efficiency of retention may even be higher, since not all complexes at this site may been stabilized by incubation with GTP. A lowered capacity
: 5 10 RNA polymerase/fd RF(mol/rnol)
Fig. 2. Isolation of G TP-stabilized RNA-polymerase . p D N A complexes. RNA-polymerase . fd-RF complexes were formed as described in the legend to Fig.l. The R F concentration was 0.58 pmol (5 x lo5 counts 32P/min) in 75 pl. After incubation with or without triphosphates (-0, A-A) GTP (0-0) 25-pl samples were digested with 25 pl DNase solution (100 pg/ml 0.5 mM CaCl,) for 10 min at 37 “C, and filtered in buffer B through nitrocellulose filters (0-0) or diluted with 0.5 ml 200 mM KCI 30 pg/ml denatured calf thymus DNA, buffer E A-A) kept for 10 min at 0 “C and filtered (0-0,
+
+
+
Eur. J. Biochem. 56 (1975)
566
Stabilization of RNA-Polymerase
DNA Complexes
i ; ( / j / i ( , O / / J ~ J / I . / ~ ~ [ , ~ ~ / , s ~ , - / ~ ~ ( ~DIVA. / ~ ~ ~ / ( , ~DNA / protected i n (A) GTP-stabilized complesca 0'-= 5 ) . (U) RNAFig. 3. ~ ~ / ; , ~ ( ~ / J l , ~ i ~ ~/J(///?UI.\ polymerase . I'd-RF complexes ( I ' = 5). (C) complexes not stabilized with G T P at I' = 10 (not retained by nitrocellulose filters after incubation at 0 C ) . (11) complexes stabilized with ATP + UTP ( r = lo), was isolated as described in Materials and Methods. The fd RF used was labelled i n the plus strand with 3zP. pDNA was depurinised and the resulting pyrimidine tracts separated by two-dimensional thin-layer chromatography as described [5]
(,30-70":,) for GTP stabilization (and for KNA chain initiation) has been observed with aged preparations of RNA polymerase, although the enzyme appeared to be normal with respect to specific DNA binding. Despite the high efficiency of the binding assay, about two enzyme molecules were always required to reach the plateau of GTP-stabilized complexes, i.e. to complex one binding site. This result was obtained with all (four) enzyme preparations used. It could be caused by (a) a partially inactive enzyme, (b) an overestimation of the protein concentration, or (c) binding of the RNA polymerase in its dimeric form. The latter explanation seems to be less likely, since isolated binding-site . polymerase complexes have been shown to contain the enzyme monomer (U. Jellinghaus, unpublished results). Isolniiori of ProtecttJd D N A ,from Stabilized Complexes
The selective stabilization of binding-site . polymerase complexes offers several possibilities for the
purification of binding sites. For the preparative isolation of the DNA from GTP-stabilized complexes, complexes were formed as described in the legends to Table 1 and Fig. 1 ( r = 5). After incubation with GTP and then at 0 "C the replicative form DNA was digested with DNase and the protected DNA was isolated as described in Methods. The pyrimidine tract analysis (Fig. 3 A) indicates the presence of only the oligopyrimidines characteristic for binding site I [5,7]. The pattern of a pDNA preparation isolated in a parallel experiment ( r = 5 , no incubation with GTP, no incubation at 0 'C) is shown in Fig. 3B. The fingerprint is more complex and corresponds to a mixture of several binding sites [ 5 ] . The comparison of the two fingerprints shows that all complexes formed at these sites dissociated during the incubation at 0 " C , except for the one formed at binding site I . The protocol of the experiment described in Fig. 2 allows the separation of binding site I from the other binding sites and thus the isolation of sites whose complexes are not stabilized with GTP. For this Cur. J . Biocheni. .i6 (1975)
C. Niisslein and H. Schaller
purpose the DNase digestion was done before the incubation at 0 "C. The DNase was inactivated by addition of EDTA and non-stabilized complexes were dissociated by incubation at 0 "C in 200 mM KCl. The resulting mixture of pDNA and pDNA . RNApolymerase complexes was fractionated by nitrocellulose filtration. The complexes remained on the filter while the pDNA appeared in the filtrate. The pyrimidine pattern is shown in Fig. 3C. As expected, the spots characteristic for binding site I (namely TC3, T3C2 and T5) [5,7] are missing. Binding site I can be eluted from the filter. In an attempt to extend the method of selective stabilization of binding-site . polymerase complexes for sites other than binding site I, we isolated complexes which had been stabilized by the combination of ATP and UTP. The pattern of this DNA (Fig. 3 D) is similar to the one shown in Fig. 3C, in that again the spots characteristic for binding site I [5,7] are missing. The complexity of the pattern, however, indicates the presence of more than one site in this preparation, suggesting that two ATP-starting complexes have been stabilized. Since the concentration of ATP and UTP required for an efficient stabilization is rather high (100 pM ATP and 200 pM UTP are required to stabilize 90% of the complexes in 10 min at 37 "C) perhaps DNA fragments from RNA synthesizing complexes might also be enriched in this preparation, resulting in an unspecific background. Whether poly(A, U) is synthesized in complexes stabilized by ATP and UTP, and if so how long such chains must be to result in stabilization, is still being investigated.
Mechanism of Stabilization with GTP The finding that only one out of four RNApolymerase . binding-site complexes can be stabilized by incubation with a single ribonucleoside triphosphate makes it highly improbable that the stabilization is due to the formation of "initiation complexes" with an altered conformation, as has been proposed by Anthony et al. [S]. The experiments to be described in this section suggest instead that the stability of the complexes is caused by the synthesis of a short oligo(G) chain within the RNA-polymerase . binding-site complex. GTP-stabilized complexes behave similarly to RNA-synthesizing complexes, which are also stable at high ionic strength and low temperature (see Table 1, last line, and [12]). The stabilization is prevented under conditions which also prevent synthesis of RNA and oligo(G) but which do not impair the binding reaction, e.g. preincubation with rifampicin or incubation without Mg2+ ions [3]. Complexes Eur. J. Biochem. 56 (1975)
567
Time ( r n i n ) Fig. 4. Incorj~oratioiz of / ' H J C ' T P iriro RNA-polymerase .fd-RF complexes. 2.5 pmol 32P-labelledfd RF (7000 counts min-' pmol- ') were incubated with 4.2 pmol RNA polymerase in 100 pI binding buffer containing 200 pg bovine serum albumin/ml and 120 mM KCI for 10 min at 37 "C. 20 pI of a [3H]GTP solution in the same buffer (60 pM GTP, 7600 counts min-' pmol -') was added and further incubated at 37 'C. 6-pl samples were taken, diluted in 0.5 ml buffer E containing 200 mM KC1, 10 pM GTP, 3 pg/ml calf thymus D N A at 0 "C. After 10 min at 0 "C the samples were filtered through nitrocellulose filters. A background of 170 counts 3H/min and 70 counts 32P/minwas substracted. (0-0) [32P]RF; (O--O) oligo([3H]G)
once stabilized, however, are insensitive to rifampicin and do not dissociate even if the Mg2+ ions are complexed by addition of excessive EDTA (data not shown). This suggests that it is the product of the reaction which renders the complexes stable. In order to detect a reaction product, labelled GTP was added to preformed RNA-polymerase . fdR F complexes. The kinetics of the stabilization in the presence of 10 pM t3H]GTP is shown in Fig. 4. In the initial period GTP label corresponding to approximately 4 GMP residues per stabilized complex is retained by the filter. After 10 min, when 90% of the complexes are stabilized, additional GMP residues are incorporated with reduced rate. One might argue that the stabilization by GTP reflects the synthesis of a normally initiated RNA product, caused by a contamination of GTP with other ribonucleoside triphosphates, rather than by the synthesis of oligo(G). Since, however, the GTP concentration required for stabilization is very low and purified GTP has the same effect, this possibility is excluded. Thus we conclude that the presence of GTP in the complex of RNA polymerase and binding site I allows an oligo(G) chain to be synthesized. This oligo(G) chain remains bound in the complex and alters its conformation, so that it no longer dissociates
568
Stabilization of RNA-Polymerase . DNA Complexes
under conditions restrictive for RNA-polymerase . DNA complexes. DISCUSSION The experiments described in the present paper show that RNA-polymerase . DNA complexes can be stabilized in special cases by preincubation with a single ribonucleoside triphosphate. It is shown that the stabilization is due to the formation of an RNA product which remains bound in a ternary complex with the enzyme and the DNA template. The same conclusion has also been drawn by McConnel and Bonner [13]who showed that apurification of the ribonucleoside triphosphates greatly reduces their ability to stabilize T7-DNA . RNA-polymerase complexes, and by So and Downey [14] who
start point (i.e. the CMP residue on the template strand coding for the starting GTP of the RNA chain) lies at the end of a C-C-C sequence approximately in the center of the binding site. This C cluster most likely acts as template for oligo(G) chain synthesis, elongation of the chain being mediated by a slippage of the product on the template [15]. This model postulates that phosphodiester bond formation can also occur next to the start point (i.e. between position - 1 and 1 instead of 1 and 2, see scheme below). Such bond formation is also indicated by the occurrence of RNA chains starting with ppp(Gp),UpA. (n = 1- 4) under apparently normal conditions for RNA synthesis (see the preceding paper [4]). Similarly RNA chain initiation next to the start point can be forced by dinucleotides complementary to the template strand, but not covering the start point [16].
+
+
+
PPPG PPPU PPPG 1 PPPG 1 Minus strand 3‘ . . . G-T-C-C-C-A-T-T . . . . . . G-T-C-C-C-A-T-T . . . 5’ -4 -3-2 -1 1 2 3 4 -4 -3-2 -1 1 2 3 4 ~~~
Initiation found that on a poly[d(A-T)] template the presence of ATP only was not sufficient to stabilize the complexes. Thus initiation complexes (defined as ternary complexes of RNA polymerase, DNA and initiating nucleoside triphosphate) do not possess a special resistance against high ionic strength, as has been speculated previously [S]. The presence of an RNA chain in the stabilized complex could alter the conformation of the DNA template by annealing to a short region of the template strand and thus stabilizing the denatured part of the DNA segment covered by RNA polymerase. This alteration of the template conformation may explain the completely altered behaviour of RNA-polymerase . binding-site complexes with respect to ionic strength and temperature. During the oligo(G) chain synthesis the polymerase apparently does not alter its position on the template since the pyrimidine tract analysis, as well as the chain length determination [3], reveal no difference between the DNA segments complexed before and after stabilization with GTP. This indicates that the oligo(G) chain synthesis takes place very close to the start point of protomer region I. The question now arises as to why, in this complex, a homopolymer synthesis takes place, while in several other complexes such a synthesis apparently does not occur. From the sequence analysis of binding site I [7] and the initiation sequence of the RNA directed by protomer region I [4], we know that the
Oligo(G) synthesis According to this explanation, stabilization by a single ribonucleoside triphosphate depends on the clustering of identical nucleotide residues at the start point. In fd replicative form DNA only complexes at one of the four sites which are stably bound by RNA polymerase at high ionic strength (120 mM KCI) can be stabilized with GTP. In addition a site which is stably bound only at low ionic strength is also stabilized under the appropriate conditions [l].No complex is stabilized with ATP or any other ribonucleoside triphosphate. Unpublished experiments indicate that complexes of RNA polymerase with diX174 replicative form DNA can be stabilized with ATP or CTP, but not with GTP. This agrees with the pppA-A-A . . . and pppC-C . . . initiation sequences observed recently in @XmRNA in vitvo [17]. Selective GTP stabilization has been also reported with a promoter region from T3 DNA [181. Thus the stabilization with single ribonucleoside triphosphates appears not to be restricted to oligo(G) synthesis and to the fd system. Although not generally applicable it provides a valuable tool for the isolation of single RNA polymerase binding sites in a pure form and high yield. Finally we may ask whether oligo(G) synthesis also occurs under physiological conditions, i.e. in the presence of the other nucleoside triphosphates. GTP stabilization, and thus also oligo(G) synthesis, is significantly reduced by the presence of UTP (Table l), probably by the incorporation of an UMP residue which prevents the “backwards slippage” of the Eur. J . Blochem. 56 (1975)
C. Nusslein and H. Schaller
newly synthesized oligonucleotide on the template. Nevertheless, even in the presence of all four nucleoside triphosphates about half of the RNA chains initiated at promotor region I were found to contain two to four G residues at their 5' terminus [4]. This result indicates that oligo(G) synthesis proceeds with a rate comparable to that of the pppG-U . . . initiation. It also shows that the oligo(G) chain produced can be utilized effectively as a primer for RNA chain growth. This interpretation is also supported by the finding that RNA chains can be initiated at promoter region I with high preference by addition of CTP, UTP, and ATP to preformed, GTP-stabilized RNApolymerase . fd-RF complexes (unpublished results).
REFERENCES 1. Seeburg, P. H. & Schaller, H. (1975) J . Mol. Biol. 92, 261277. 2. Hinkle, D. & Chamberlin, M. (1970) Cold Spring Harbor Symp. Quant. Biol. 35, 65 - 72. 3. Niisslein, C. (1973) Ph. D. Thesis, Tubingen.
569 4. Heyden, B., Nusslein, C. & Schaller, H. (1975) Eur. J . Biochern. 55, 147-155. 5. Heyden, B., Nusslein, C. & Schaller, H. (1972) Nut. New B i d . 240,9- 12. 6. Seeburg, P. H. (1975) Ph. D . Thesis, Tubingen. 7. Schaller, H., Gray, C. P. & Herrmann, K. (1975) Proc. Nut1 Acad. Sci. U.S.A. 72, 737-741. 8. Anthony, D. D., Zeszotek, E. & Goldwaith, D . A. (1966) Proc. NatlAcad. Sci. U.S.A. 56, 1026-1033. 9. Stead, N. W. & Jones, 0. W. (1967) J . Mol. Biol. 26, 131- 135. 10. Hinkle, D. & Chamberlin, M. (1972) J . Mol. Biol. 70, 157185. 11. Heyden, B. (1974) Ph. D. Thesis, Tubingen. 12. Naito, S. & Ishihama, A. (1973) Biochem. Biophys. Res. Commun. 51,323 - 330. 13. Mc Connell, D . J. & Bonner, J. (1973) Eur, J . Biochem. 38, 111 -121. 14. So, A. G. & Downey, K. M. (1970) Biochemistry, 9, 47884793. 15. Chamberlin, M. & Berg, P. (1962) Proc. Nut1 Acad. Sci. U.S.A. 48,81-94. 16. Minkley, E. G. & Pribnow, D . (1973) J . Mol. Bid. 77, 255277. 17. Smith, L. H., Grohmann, K. & Sinsheimer, R. L. (1974) Nucleic Acids Research, 1, 1521- 1529. 18. Takeya, T. & Takanami, M. (1974) Biochemistry, 13, 53885394.
C. Nusslein, Biozentrum der Universitat Basel, Klingelbergstrasse 70, CH-4056 Basel, Switzerland
H. Schaller *, Mikrobiologie der Ruprecht-Karl-Universitat Heidelberg, D-6900 Heidelberg 1, Im Neuenheimer Feld 280, Federal Republic of Germany
* To whom correspondance should be sent
Eur. J. Biochem. 56 (1975)
