Initiation and Release of RNA by DNA-Dependent RNA Polymerase* Robert L

Proceedings of the National Academy of Sciences Vol. 66, No. 3, pp. 701-708, July 1970

Initiation and Release of RNA by DNA-Dependent RNA Polymerase* Robert L. Millettet and Carol D. Trotter

DEPARTMENT OF PATHOLOGY, UNIVERSITY OF COLORADO MEDICAL CENTER, DENVER Communicated by James Bonner, December 19, 1969 Abstract. During in vitro transcription of T4 DNA by E. coli RNA polymerase, chain initiation stops coincidentally with synthesis at low ionic strength (0.11) with an average of one RNA chain initiated per 24S polymerase molecule. At high ionic strength (0.37), initiation as well as synthesis continues for several hours, with an average of four chains initiated per enzyme molecule in two hours. The RNA product is released from the T4 DNA template at both low and high ionic strength. At high ionic strength, however, RNA polymerase can repeatedly initiate, synthesize, and release RNA from the synthesis complex in vitro. Under both conditions, the synthesized RNA sediments at 25-44 S, has a number average chain length of 5000 to 7500 nucleotides, and a weight average chain length of 11,500 nucleotides.

Introduction. We previously demonstrated that the cessation of RNA synthesis in vitro can be overcome by elevating the salt concentration.1 Thus, at an ionic strength of 0.37, RNA synthesis continues over many hours with only slightly decreasing rate. Preliminary experiments based on the incorporation of ly-32P-ATP indicated that high salt concentrations allow RNA polymerase to initiate RNA chains continuously.1 These observations suggested that the polymerase is able to terminate RNA chains under these conditions. We show here that at high ionic strength RNA polymerase continually initiates and releases RNA of high molecular weight from the synthesis complex in vitro. In addition, we find that even at low ionic strength, where synthesis terminates by 1 hr, the bulk of the product is released. This work has been presented in preliminary form elsewhere.2 Materials and Methods. RNA polymerase was prepared from 250 g Escherichia coli K12 (3/ log phase, washed, from Grain Processing Corp.) as previously described1 but with the following modifications: The DNA-protein complex was precipitated from the crude extract with a 1-5% w/v solution of polyethylenimine (Polysciences, Inc.)4 in TMA buffer' containing 2.5 X 10-4 M EDTA. An amount determined to precipitate 95% of the polymerase activity was added slowly with stirring to the extract. The result- ing precipitate was centrifuged and washed with 195 ml of 0.2 M NH4Cl in TMA, and the enzyme eluted by homogenizing with 170 ml of 0.4 M NH4Cl in TMA and centrifuging. After ammonium sulfate fractionation of the eluate,3 the polymerase was chromatographed on a 3 X 33 cm DEAE-cellulose column as previously described.3 Using precipitation with 55% saturated (NH4)2S04 to concentrate the enzyme between steps, we further purified it by two zone centrifugations: the first at low ionic strength (10-30% w/v sucrose in TMA),3 and the second at high ionic strength (10-30% w/v sucrose in 0.4 M NH4Cl in TMA, 25,000 rpm for 27 hr at 20C in the SW-25.1 rotor) to separate the en- 701 Downloaded by guest on September 30, 2021 702 BIOCHEMISTRY: MILLETTE AND TROTTER PROC. N. A. S.

zyme in the 13S form from larger contaminants. The precipitated enzyme was dissolved in 0.04 M potassium phosphate, 1 mM f3-mercaptoethanol, 1 mM EDTA, pH 7.0, dialyzed for 2 hr against the same buffer, and then chromatographed on hydroxylapatite according to the method of Richardson,5 but with 1 mM EDTA present in the chromatography buffer. Two peaks of polymerase activity elute from the hydroxylapatite column, one at about 0.13 M phosphate, the other at about 0.16 M phosphate. Each peak was precipitated with ammonium sulfate, dissolved in glycerol-2X TMA (1:1), pH 7.9, and stored at -20'C. The first eluting enzyme peak, which contained subunits B, O', a, a, and co6 was used in these studies. Its specific activity averaged 1800 enzyme units9/mg protein (determined from a specific absorbance at 280 mu of 0.65/mg/ml).' RNA synthesis was measured as published earlier' at either low ionic strength, 1A = 0.11 (50 mM NH4Cl, 8 mM magnesium acetate, 30 mM Tris, pH 7.9), or at high ionic strength, A = 0.37 (0.13 M NH4Cl, 0.07 M magnesium acetate, 0.03 M Tris, pH 7.9), with either UTP-3H (0.5 or 1 Ci/mole) or ATP-8-'4C (1 Ci/mole) used as the labeled substrate. Chain initiation was followed by the incorporation of y-32P-ATP or -GTP7 having specific activities of 26-58 cpm/pmole. To eliminate any artifacts owing to possible polyphosphate kinase contamination, 0.02 mM ADP was included in the reaction mixture.8 Aliquots of 0.1 ml were pipetted into 1 ml ice-cold H20, then 0.05 ml unlabeled ATP (100 ,umoles/ml) plus 1 ml 0.12 M sodium pyrophosphate (pH 6-7) were added with mixing. 2.2 ml 10% trichloroacetic acid containing 0.02 M sodium pyrophosphate were added with mixing and the samples were filtered through glass filters (Whatman GF/C) presoaked in 5% trichloroacetic acid containing 0.1 M pyrophosphate. Filters were washed five times with 7 ml of 5% trichloroacetic acid-0.1 M pyrophosphate, three times with 5% trichloroacetic acid, and twice with ethanol, dried, and counted in a liquid scintillation counter. Determination of s20,W values and molecular weights of RNA from sucrose gradients: Zone centrifugation procedure is described in the legends. Individual gradients were calibrated by computer analysis to give s20o, as a function of r, with R17 phage and R17 RNA as internal standards (J. McConnell, to be published). Sedimenta- tion coefficients were converted to molecular weights using Spirin's equation,9 and molecular weight distributions were determined for each gradient. Template DNA: 3H-Thymidine-labeled T4 phage were prepared by the method of Richardson et al.'0 T4 phage were purified by sedimentation through sucrose gradients," dialyzed against NET (0.1 M NaCl, 1 mM EDTA, 0.01 M Tris, pH 7.5), and the DNA prepared by extracting with phenol and chloroform-octanol.' The free T4 DNA had an 820,w of 46 S in our sucrose gradients. 3H-R17 phage RNA: 3H-Labeled R17 phage RNA was prepared by the method of Strauss and Sinsheimer12 using, as host, E. coli Hfr Hayes (AB259, obtained from A. L. Taylor). Three min after infection 0.8 mCi of uridine-5-3H (20 Ci/mmole) was added to a 1 liter culture. Chemicals: y-32P-ATP and -GTP were synthesized enzymatically by the method of Glynn and Chappell"3 with a modification after the Dowex-1 chromatography step to yield the lithium salts.'4 Unlabeled nucleotides were obtained from P-L Biochemicals, Inc., 3H- and "4C-labeled substrates from Schwarz BioResearch, and hydroxylapatite from Bio-Rad. Ammonium sulfate (enzyme grade) and sucrose (enzyme grade, for polymerase purification; density gradient grade, for zone centrifugation of the reaction product) were purchased from Mann Research Labs., and DEAE-cellulose (Type 20) from Schleicher & Schuell. Results. Kinetics of chain initiation and synthesis at low and high ionic strength: In vitro transcription under the usual low ionic strength conditions (u = 0.11) terminates in about 30 min-." The stopping of synthesis at low ionic strength is accompanied by a parallel cessation in chain initiation as evi- Downloaded by guest on September 30, 2021 VOL. 66, 1970 BIOCHEMISTRY: MILLETTE AND TROTTER 703

50 SO p = 0.37

40 * ,L- 0.11 FIG. 1.-Kinetics of chain initiation and synthesis at low and high ionic strength. RNA synthesis was performed at .i low ionic strength, ju = 0.11, and > 30 15 at high ionic strength, s =/ 0.37, with 11 ug RNA poly- X/ merase/ml and 98 ug T4 DNA/ ml. Labeled substrates: 3H-/ UTP, 0.5 Ci/mole, oy..2P-ATP E 20 - 1 + (31 cpm/pmole) and _y-32P-GTP / / X (33.2 cpm/pmole). Duplicate reactions were run containing either 3H-UTP and 32P-ATP or 3H-UTP and 32P-GTP. 0.2- ml aliquots taken at various 10 -5

times were assayed as described 0 in Materials and Methods. QE

0 30 60 90 120 150 Ml nutes denced by the incorporation of y-32P-ATP and -GTP (Fig. 1). Similar findings have teen reported previously.7" 6 Synthesis at high ionic strength (0.37), however, continues for several hours with only slightly decreasing rate as we have earlier demonstrated.' Under these conditions, chain initiation shows a rapid initial rise during the first 10 min and then increases almost linearly throughout the course of transcription (Fig. 1). By 2 hr of synthesis, approximately three times as many RNA chains are initiated at high ionic strength as at low ionic strength. The ratio of ATP/GTP initiation is unaffected by the ionic conditions used in these experiments: At both low and high ionic strength, 67% of the RNA chains transcribed from T4 DNA are initiated with ATP. This agrees with values reported for initiation at low salt concentrations.8 Effect of ionic strength on chain length and number of RNA molecules synthesized per enzyme particle: From an average of five experiments, the number of TABLE 1. Effect of ionic strength on RNA chain length and number. Number of RNA Chains* Number-Averaget per Polymerase Molecule Chain Length (Nucleotides) Minutes , = 0.11 0.37 0.11 0.37 5 0.654 0.316 4335 4592 15 0.973 0.934 5274 5507 30 0.928 1.58 6116 7059 60 1.09 2.72 5116 7761 90 1.22 3.62 6311 7574 120 1.38 4.23 5413 7678 * Calculated from nmoles 7-32P-ATP + y-32P-GTP incorporated/nmoles enzyme; nmoles enzyme = ng enzyme/720,000. t Calculated from nmoles UMP incor)orated/0.33 X inmoles (y-32P-ATP + 7-32P-GTP) incorporated. 0.33 = mole fraction of A in T4 DNA. jug DNA/jg polymerase ranged from 1.0 to 5.7. Downloaded by guest on September 30, 2021 704 BIOCHEMISTRY: MILLETTE AND TROTTER PROC. N. A. S.

RNA chains synthesized per 24S polymerase molecule (M = 720,000)17 has been calculated (Table 1). At low salt concentrations, each polymerase molecule initiates approximately one RNA 15 mi chain by the time synthesis terminates. 4 _ k ) _ 15 At elevated ionic strength, however, the number of chains initiated per enzyme l i 10 particle continually increases with syn- 2 thesis time so that by 2 hr each polymerase molecule has initiated, on the average, about four RNA chains. .\taft/ 9 _ From the ratio of 3H-UM\'P to y_32P_ l ATP and -GTP incorporation in the 30 mn x same set of experiments, the number o average chain length of the product has 4 been calculated (Table 1). The maxi- c -15 mum chain length is reached by 30 min of synthesis. At lowsalt concentrations 2 1110 ; the average chain length after termina- <2 - al \ o tion of synthesis is approximately 5700 0 nucleotides; at high ionic strength this E a. quantity reaches a steady-state level of r j 2 I>about 7500 nucleotides. Multicistronic I1'uRNA's are thus produced under both 60 min salt conditions. From the average chain lengths at 5 4 | min of synthesis, average chain growth 15asrates of 14.4 and 15.3 nucleotides/sec have been calculated for the low and 10 high ionic strength conditions, respec- 2 tively. These are in good agreement 5 with the value of 16 nucleotides/sec reported by Richardson.18 Since initia- ; < > tion is asynchronous, these probably 10 20 are minimal values. Fraction no. Release of product from the synthe- FIG. 2.-Sucrose gradient sedimentation of sis complex: The above observations a high ionic strength reaction mixture. RNA at high ionic strength suggested that synthesis was performed at high ionic strength such conditions promote the release of (0.37) with 9 jug polymerase/ml, 123 tg the RNA product. To verifythis, RNA (132,000 cpm) 3H-T4 DNA/mi, and with 14C-ATP, 2 Ci/mole, as labeled substrate. was transcribed from 3H-labeled T4 50 IAl samples were taken at the indicated DNA with 14C-ATP as the labeled sub- times, diluted 20X into NET (see Materials strate. At various times of synthesis, and Methods). 0.3 ml of the diluted samples were layered onto 12.1 ml sucrose gradients aliquots of the reaction mixture were di- (5-30% w/v in NET) and centrifuged for luted 20-fold and centrifuged on sucrose 1.68 hr at 41,000 rpm at 30C in the SW-41 gradients (Fig. 2). At early times (0- rotor. Sedimentation is from right to left. Fraction zero represents the pellet fraction. 15 min) most of the product is found in Downloaded by guest on September 30, 2021 VOL. 66, 1970 BIOCHEMISTRY: MILLETTE AND TROTTER 705

association with the template, presumably in a ternary complex with the enzyme as initially shown by Bremer and Konrad.'5 With increasing time of synthesis, however, one sees an increasing amount of free RNA sedimenting faster than the T4 DNA template. By 2 hr of synthesis essentially all of the product sediments at least twice as fast as the template and accumulates in the pellet fraction (Fig. 3, top). This is a result of aggregation as is shown by dissociation of the RNA to a lighter sedimentation distribution by heating the diluted incubation mixture at 450C for 5 min before centrifugation (Fig. 3, bottom).

26.6 2h synthesis T44 DNA

-6 T4 10min x DNA

,2 6'_PhageE - 4 ~~~~~~~~~~~~R17

-- 0 e E

gran heated 5 min 45wC 2m I 6<

C1 6~~~~~~~~U be 0 minI 3

3;Z

0 r 10 20 Fraction no. 2,o etiuaint41,000rm0m z2 0 ~~~~~~~~~~~~~~0 (Above) FIG. 3.--Sucrose gradient analy- 0 sis of2i a lo2-hr oisynthesisteghratinmxuemixture at high ionic strength. RNA synthesis and sucrose gradient centrifugation were performed as in Fig. 2, but with 2 hr of synthesis and 2.33 hr of centrifugation at 41,000 rpm.me6 Top: 20X diluted aliquot applied directly to gradient; Bottom: diluted sample was heated 5 minh at 4500 and quick cooled 4 3 before layering. 2.3 (Right) FIG. 4.-Sucrose gradient analysis of a low ionic strength reaction mixture. 2 Synthesis conditions were as in Fig. 2, but at low ionic strength (0.11). Samples 2 were diluted as in Fig. 2, and centrifuged1 for 2.33 hr at 42,000 rpm. 6 X 1010 plaque- forming units of phage R17 (79 5) were added to the 10-mins gradient as sedimentation marker. The T4 DNA peak had an I10 20 820,w Of 47 5. Fraction no. Downloaded by guest on September 30, 2021 706 BIOCHEMISTRY: MILLETTE AND TROTTER PROC. N. A. S.

That the aggregate is composed of free RNA and not a DNA-RNA complex is shown by the following: (a) The percentage of DNA banding in the major DNA peak (75.8%o, average), as well as that sedimenting to the bottom of the gradient (12.8%, average), remains relatively constant with increasing synthesis time. The rapid sedimentation of a fraction of T4 DNA to the bottom of sucrose gradients is a previously observed phenomenon.19 (b) From 15 to 60 min of synthesis the percentage of '4C-RNA in the DNA region decreases from 69 to 37%0 while in the pellet and three bottom fractions it increases from 12 to 24%0. (c) Shifting the ionic strength of a 90-min synthesis mixture from 0.11 to 0.37 immediately before sucrose gradient analysis induces the RNA aggregation but does not affect the DNA distribution. (d) The purified RNA (phenol-extracted from a high ionic strength reaction mixture), when not heated or dialyzed against a low ionic strength buffer containing EDTA, shows a similar aggregation behavior. The aggregation is not a result of extensive regions of self-complementarity because (a) only 5%0 of the purified RNA remains RNase resistant both before and after heating to 70'C and quick cooling, and (b) the purified RNA is not dissociated on DMSO (dimethylsulfoxide) gradients (5-20%o w/w sucrose in 99% DMSO, 10-3 lI EDTA). These and the above observations suggest that the RNA product aggregates as a result of the high ionic strength and high Mlg2+ concentration.20 Diluted reaction mixtures from a low ionic strength synthesis mixture were similarly analyzed on sucrose gradients. Here we expected to see little or no release of product as reported previously.15'2' However, as demonstrated in Figure 4, release of the product is already apparent by 10 min of synthesis. The product shows little aggregation, in this case, and accumulates mainly in a position lighter than the template. By 60 min of synthesis, RNA sedimenting with peaks at 28 and 41 S is observed. The percentage of total '4C counts sedimenting above the 'H-T4 DNA gives a conservative estimate of the released RNA. This value, from six experiments, is 47%0 after 60 min of synthesis. However, diluting the reaction mixture into 0.2% sodium dodecyl sulfate in NET before sucrose gradient sedimentation, which is known to release RNA from the enzyme-DNA complex,'5 causes no significant change in the RNA profile at 60 min of synthesis. Therefore, by the time synthesis terminates at low ionic strength, most of the product is free from the template. Size of the released RNA: From an average of five experiments at low ionic strength, the sedimentation constants of the two main peaks of the released RNA product were 25 and 44 S. From the sedimentation distributions on the same gradients, molecular weight distributions were determined and chain lengths of 11,400 nucleotides, weight average, and 6100 nucleotides, number average, were calculated. An evaluation of several experiments at high ionic strength gave similar results: the RNA peaks ranged from 27 to 43 S, while the weight- and number-average chain lengths, as calculated from the total distribution, were 11,600 and 4800 nucleotides, respectively. Release is not due to ribonuclease: In order to demonstrate that the observed release of RNA is not due to ribonuclease contamination of the reaction Downloaded by guest on September 30, 2021 VOL. 66, 1970 BIOCHEMISTRY: MILLETTE AND TROTTER 707

mixture, 3H-labeled R17 phage RNA, sedimenting at 27 S, was added to a low ionic strength synthesis mixture 5 min after the start of incubation at 370C. The time delay was invoked to avoid possible inhibition of uninitiated polymerase by the added RNA. Incubation was continued at 370C and samples, which were taken at various times, were analyzed on sucrose gradients. After 3 hr of incubation, there was no significant alteration in the sedimentation constant or profile of the added R17 RNA. Discussion. We have demonstrated that at high ionic strength RNA polymerase can repeatedly initiate, synthesize, and release RNA from the synthesis complex in vitro. The released product, as well as the purified RNA, forms a fast sedimenting aggregate, presumably due to the high salt concentration.20 It is dissociable by a brief heating at 450C or by dialysis against EDTA. Our finding of product release at high ionic strength agrees with a report by Maitra and Barash,22 using similar techniques, and experiments by Richardson,'8 using a membrane filter assay. We were surprised to find that a major portion of the product is also released from the template during synthesis at low ionic strength. This contradicts the earlier findings of Bremer and Konrad'5 which show the retention of product in the enzyme-template complex at the completion of synthesis. Possible explana- tions for these differences are: (a) differences in the salt composition of the reaction mixtures, (b) differences in the enzyme preparations used, and (c) free RNA of the size we have observed would cosediment with the broad T4 DNA band in the gradients of Bremer and Konrad. From the double-labeling experiments we calculated number-average chain lengths of about 5700 and 7500 nucleotides for low and high salt incubations respectively. These figures might be somewhat in error since the calculations are based on the base composition of T4 DNA. From the sedimentation distributions on sucrose gradients, RNA chains of comparable size are produced at both low and high ionic strength (both are around 11,500 nucleotides, weight average). Errors inherent in the Spirin formula, especially at very low and high S values, may cause inaccuracies here. However, this value is in excellent agreement with the average size calculated for transcription regions in the early genes of T4: From studies on the effect of ultraviolet irradiation of T4 phage on in vivo transcription, we have found these regions to be about 10,000 base pairs in length.24 Under our low salt assay conditions, each polymerase molecule initiates only one chain, on the average, with most of the RNA being released from the DNA. These findings, along with the fact that synthesis stops as a result of product inhibition,25 indicate that at low ionic strength the enzyme becomes inhibited, after one complete round of transcription, by the released RNA chain. The fact that almost identical weight-average chain lengths are synthesized at both low and high ionic strength gives added support to this interpretation. It therefore seems likely that the previously reported first-order inactivation of the enzyme' may be simply the consequence of the kinetics of chain initiation, because, as evidenced by the y-32P-ATP incorporation, it takes approximately 15 min for all enzyme molecules to initiate at low ionic strength. Downloaded by guest on September 30, 2021 708 BIOCHEMISTRY: MILLETTE AND TROTTER PROC. N. A. S.

The experiments presented have shown that RNA polymerase can repeatedly initiate, synthesize, terminate, and release RNA in vitro. Is the release just a random process, or is it a specific release of completed mRNA molecules at prescribed termination sites on the template? The latter alternative is supported by the observation that the chain length of in vitro-synthesized T4 RNA agrees closely with the calculated size of T4 transcription regions in Vivo.24 We are hopeful that a definitive answer to this question will come from further charac- terization of the released product by polyacrylamide gel electrophoresis and 3'-end analysis. Note added in proof: Reinitiation by RNA polymerase at high ionic strength has been reported recently by J. P. Richardson in Nature, 225, 1109 (1970). We would like to thank Drs. D. Pettijohn and W. Sauerbier for their helpful discussions and reading the manuscript. Special thanks are due J. McConnell for his computer analysis of the sucrose gradient data. * This work was supported by grants from the U.S. Public Health Service, CA 05164, and National Science Foundation, GB-7818. t Requests for reprints may be addressed to Dr. R. L. Millette, Department of Pathology, University of Colorado Medical Center, Denver, Colo. 80220. 1 Fuchs, E., R. L. Millette, W. Zillig, and G. Walter, Europ. J. Biochem., 3, 183 (1967). 2 Millette, R. L., Federation Proc., 28, 659 (1969), abstract. 3 Zillig, W., E. Fuchs, and R. L. Millette, in Procedures in Nucleic Acid Research, ed. G. L. Cantoni and D. R. Davies (New York: Harper and Row, 1966), p. 323. Zillig, W., K. Zechel, and H. J. Halbwachs, Z. Physiol. Chem., 351, 221 (1970). Richardson, J. P., these PROCEEDINGS, 55, 1616 (1966). 6 Burgess, R. R., A. A. Travers, J. J. Dunn, and E. K. F. Bautz, Nature, 221, 43 (1969). 7Maitra, U., and J. Hurwitz, these PROCEEDINGS, 54, 815 (1965). 8 Maitra, U., Y. Nakata, and J. Hurwitz, J. Biol. Chem., 242, 4908 (1967). 9 Spirin, A. S., Biokhimiya, 26, 511 (1961). 10 Richardson, C. C., R. B. Inman, and A. Kornberg, J. Mol. Biol., 9, 46 (1964). 11 Pettijohn, D., Europ. J. Biochem., 3, 25 (1967). 12 Strauss, J. H., and R. L. Sinsheimer, J. Mol. Biol., 7, 43 (1963). 13 Glynn, I. M., and J. B. Chappell, Biochem. J., 90, 147 (1964). 14Smith, M., and H. G. Khorana, J. Amer. Chem. Soc., 80, 1141 (1958). 15 Bremer, H., and M. W. Konrad, these PROCEEDINGS, 51, 801 (1964). 16 Bremer, H., M. W. Konrad, K. Gaines, and G. S. Stent, J. Mol. Biol., 13, 540 (1965). 17 Priess, H., and W. Zillig, Biochem. Biophys. Acta, 140, 540 (1967). 18 Richardson, J. P., in Progress in Nucleic Acid Research, ed. J. N. Davidson and W. E. Cohn (New York: Academic Press, 1969), vol. 9, p. 75. 19 Rosenbloom, J., and V. N. Schumaker, Biochemistry, 2, 1206 (1963). 20 Boedtker, H., in Methods of Enzymology, ed. L. Grossman and K. Moldave (New York: Academic Press, 1968), vol. 12B, p. 433. 21 Richardson, J. P., J. Mol. Biol., 21, 115 (1966). 22 Maitra, U., and F. Barash, these PROCEEDINGS, 64, 779 (1969). 23 Geiduschek, E. P., L. Snyder, A. J. E. Colvill, and M. Sarnat, J. Mol. Biol., 19, 541 (1966). 24 Sauerbier, W., R. L. Millette, and P. Hackett, Biochim. Biophys. Acta, in press. 2 Krakow, J. S., J. Biol. Chem., 241, 1830 (1966). Downloaded by guest on September 30, 2021