Eur. J. Biochem. 73, 179-184(1977)

Transcription of Double-Stranded RNA by Escherichia coli DNA-Dependent RNA Polymerase Masahiro SUGIURA and Kin-ichiro MIURA National Institute of Genetics. Mishima (Received September 16/November 15, 1976)

Double-stranded RNA of some virus genomes can be used as template for the DNA-dependent RNA polymerase purified from Escherichia coli. The RNA synthesis requires all four nucleoside triphosphates and manganese ions and is dependent on the presence of sigma subunit. The reaction is inhibited by rifampicin, streptolydigin and ethidium bromide, but not by DNase and actinomycin D which does not bind to double-stranded RNA. The template activity of double-stranded RNA from various viruses is different in each case. The order of template efficiency is Penicillum chrysogenum virus > cytoplasmic polyhedrosis virus > rice dwarf virus > reovirus. The product obtained using cytoplasmic polyhedrosis virus double-stranded RNA as template is single-stranded and hybridizes specifically to the denatured template RNA. One of the major 5'-starting nucleotide sequences of the product RNA is pppA-A-Y---. These results indicate that transcription in vitro of double-stranded RNA by E. coli RNA polymerase is initiated at specific sites on the template. Bacterial DNA-dependent RNA polymerases can utilize a variety of synthetic polyribonucleotides as effective templates and can synthesize their complementary polymers [l - 101. Naturally occurring single-stranded RNA also supports RNA synthesis catalyzed by the bacterial enzymes, although to a lesser extent than DNA and synthetic ribopolymers [l - 31. Contradictory reports, however, have appeared in the literature concerning the template activity of the dsRNA obtained from diplornaviruses which contain dsRNA as genetic materials. For example, Gomatos et al. reported that the viral dsRNA was an effective template for a bacterial RNA polymerase [ll], while Shatkin [12] and Miura et al. [13] suspected the template activity of the dsRNA. Before using viral dsRNA as a possible model template for Escherichia coli DNA-dependent RNA polymerase, it was necessary to reinvestigate carefully the template activity of the dsRNA. In this report, we present evidence that the viral dsRNA does serve as template for E. coli RNA polymerase. Particular attention has been paid to examining the ability of E. coli enzyme to recognize the specific regions on the dsRNA where transcription is initiated. Abbreviations. dsRNA, double-stranded RNA ; CPV, cytoplasmic polyhedrosis virus. Enzymes. DNA-dependent RNA polymerase (EC 2.7.7.6); pancreatic RNase (EC 3.1.4.22); RNase TI (EC 3.1.4.8); DNase (EC 3.1.4.5).

MATERIALS A N D METHODS Preparation of RNA Polymerase

RNA polymerase was prepared from E. coli A19 (RNase I-) by the method outlined by Burgess and Travers [14]. The glycerol gradient step was replaced by two cycles of the high and low salt agarose gel filtration. The enzyme was further purified by a T4 DNA-cellulose column as described by Bautz and Dunn [15]. Electrophoresis on dodecylsulfate-polyacrylamide gel indicated that the enzyme preparation was at least 95 % pure and contained about 0.6 equivalent CJ subunit. It had a specific activity of 600 units/mg protein with T4 DNA as template, where one unit was equivalent to 1 nmol UMP incorporation for 10 min at 37 "C. The purified enzyme had no detectable RNase activity; incubation of 10 pg of the enzyme with 0.4 pg of E. coli [3H]mRNA at 37 "C for 7 h did not produce any acid-soluble materials. The enzyme preparation was essentially free of polynucleotide phosphorylase activity when examined under the conditions for RNA synthesis [16]. When 10 pg of the RNA polymerase was incubated at 37 "C for 20 min in 0.2 ml of 0.04 M Tris-HC1 (pH 7.9 at 25 "C), 3 mM MnCl,, 0.1 mM dithiothreitol, 0.1 mM EDTA and 0.2 mM [3H]UTP, less than 0.008 nmol of 13H]UMP incorporation into acid-insoluble materials was observed. Core enzyme and CJ subunit were prepared by

Double-Stranded RNA Transcription

180

phosphocellulose column chromatography according to Burgess et al. [17]. The enzyme was stored at -20 "C at a concentration of 1 mg/ml in 50 % glycerol, 0.01 M Tris-HCI (pH 7.9), 0.01 M MgCI,, 0.1 M KCI, 0.1 mM dithiothreitol and 0.1 mM EDTA. Preparation of Templates

dsRNA from cytoplasmic polyhedrosis virus (CPV) was prepared as described previously [18]. Rice dwarf virus was prepared by a modified procedure of Kimura and Kodama [19]. To 1 1 of leaf-extract were added 13.4 ml of 20% dextran sulfate 500, 290 ml of 30% poly(ethylene glycol) 2000 and 50 ml of 5 M NaCl and allowed to settle. The viral particles were collected by high-speed centrifugation from the lower phase and interface after removal of dextran sulfate by KCl. The dsRNA was extracted from the rice dwarf virus preparation as described previously [20]. Reovirus and Penicillium chrysogenum dsRNAs were kind gifts of Dr F. Koide and Dr K. Yazaki respectively. E. coli 23-S rRNA was prepared as described [21]. T4 phage DNA was prepared according to Thomas and Abelson [22]. Poly(A) and calf thymus DNA were purchased from Boehringer Mannheim. Conditions for R N A Synthesis

The standard reaction mixture contained a total volume of 0.2 ml: 0.04 M Tris-HC1 (pH 7.9 at 25 "C), 3 mM MnCl,, 0.1 mM dithiothreitol, 0.1 mM EDTA, 0.2 mM each of ATP, CTP, GTP and rH]UTP (25 Ci/mol, Radiochemical Center), 10 pg CPV dsRNA and 10 pg RNA polymerase. After incubation at 37 "C for 20 min, the reaction was terminated by adding 100 pg of carrier RNA followed by 0.25 ml of cold 20 % trichloroacetic acid containing 0.02 M PP,. The acid-insoluble material was collected on a glass filter and counted. Rates of RNA synthesis were expressed as nmol [3 HIUMP incorporated into acid-insoluble materials, after subtraction of the value obtained in a control reaction lacking template (0.006 nmol). Analysis of the R N A Products

Total volume of the standard reaction mixture, specific activity of [3 HIUTP and incubation period were increased to 0.4 - 1.O ml, 125 Ci/mol and 40 min, respectively. To label RNA with [ Y - ~ ~ P I A T0.2 P , mM [p3,P]ATP (1.5 x lo7 dis. min-' nmol-') and 0.1 mM ADP, an inhibitor of polyphosphate kinase [16], were included. After incubation at 37 "C for 40 min, the reaction was stopped by adding sodium dodecyl sulfate to 0.1 % and EDTA to 0.01 M. This mixture was treated twice with 80% phenol. The separated aqueous phase was then passed through a Sephadex

G-50 column (1 x 30 cm) equilibrated with 0.02 M Tris-HC1 (pH 7.4) and 0.15 M NaCl. The labeled RNA eluted at void volume was collected and precipitated by ethanol. To examine RNase sensitivity, the [3 HIRNA was dissolved in 0.01 M Tris-HC1 (pH 7.4), 0.3 M NaCl and 1 mM EDTA, and treated with a mixture of RNase A (10 pg/ml) and RNase TI (20 units/ml) for 30 min at 37 "C. Under these conditions, E. coli [3 HImRNA was completely hydrolyzed but not CPV dsRNA. To anneal, the [3H]RNA was mixed with excess of CPV dsRNA (20 pg/ml) and adjusted to 0.01 M Tris-HC1 (pH 7.4) and 1 mM EDTA. The mixture was denatured by heating at 100 "C for 10 min, added NaCl to 0.3 M and incubated at 72 "C for 18 h. After slow cooling it was treated with RNases as above. T o determine the size of the product RNA, the [3H]RNA was layered onto a 5-20% linear sucrose gradient in 0.02 M Tris-HC1 (pH 7.4) and 0.15 M NaCl, and centrifuged at 24000 rev./min for 18 h in a swinging-bucket roter SW27 on a BeckmanSpinco ultracentrifuge. For denaturation, the [3 HIRNA was dissolved in 90 % formamide containing 1.5% formaldehyde and 0.05 M NaCI, heated at 60 "C for 5 min and cooled quickly in ice [23]. Analysis of the starting nucleotide sequences of the RNA labeled with [p3,P]ATP was done as described previously [24]. Each fraction obtained by centrifugation or chromatography was dissolved in toluene/ methylcellosolve-based scintillation liquid [21] and counted. RESULTS General Characteristics of the RNA Synthesis

A highly purified DNA-dependent RNA polymerase preparation free of interfering enzyme activities was necessary because poor template activity of dsRNA was expected. The enzyme from E. coli A19, essentially free of RNase and polynucleotide phosphorylase activity was obtained by two cycles of the high and low salt agarose gel filtrations followed by T4-DNA cellulose column chromatography (see Materials and Methods). When Mg2+was replaced with Mn2+in a standard DNA-dependent RNA synthesis system catalyzed by E. coli RNA polymerase [14], dsRNA from CPV supported [3H]UMPincorporation into acid-insoluble materials. The rate of incorporation was only 1/20 of that obtained by the use of calf thymus DNA as template (Table2). As shown in Tablel, the dsRNAdependent [3 HIUMP incorporation required all four nucleoside triphosphates besides Mn2 . Rifampicin, streptolydigin, congo red and aurintricarboxylic acid, which are known to be potent inhibitors of E. coli RNA polymerase because of their interaction with +

181

M. Sugiura and K. Miura

t

Tablel. Characteristics of CPV dsRNA-dependent [ 3 H ] U M P incorporation Reaction conditions were as described in Materials and Methods with the omission or addition indicated Conditions

r3 HIUMP incorporated nmol

Complete -dsRNA -MnZ+ - ATP - CTP - GTP -ATP, CTP, GTP + Rifampicin (3 pM) + Streptolydigin (0.5 mM) +Congo red (25 pM) +Aurintricarboxylic acid (50 p M ) +RNase A (10 pg) +DNase(5 pg) +Actinomycin D (8 pM) +Ethidium bromide (0.2 mM)

0.176 0 0.01 1 0.001 0.004 0.009 0.006 0 0.002 0.002 0.005 0.002 0.162 0.164 0.016

protein, caused severe inhibition. RNase also inhibited the [3H]UMP incorporation. By contrast, the reaction was insensitive to DNase and to actinomycin D, which was shown not to interact with dsRNA [12]. Ethidium bromide, an intercalating agent which binds to both double-stranded DNA and RNA, inhibited the [3H]UMP incorporation. These results clearly showed that the observed incorporation of rH]UMP was due to the action of RNA polymerase using dsRNA as template and was not due to that of contaminating enzymes or DNA. Addition of CPVdsRNA to the DNA-dependent RNA synthesis reaction did not cause any inhibition (data not shown), indicating that the low template activity of the dsRNA was not due to the presence of an inhibitor in the dsRNA preparation. The incorporation of [3 HIUMP into acid-insoluble materials increased with the amount of dsRNA added, up to 10 pg. The [3H]UMP incorporation began with a short lag period and continued linearly for at least 40min. After incubation for 40min, 0.36nmol of [3 HIUMP was incorporated into RNA. Assuming that the product contained 25 % UMP, it corresponded to 1.44 nmol of RNA nucleotides. Thus the RNA product represented 0.05-fold increase over the dsRNA added (30 nmol nucleotides). The effect of the rifampicin concentration on the initiation of RNA synthesis when RNA-polymerase . dsRNA complexes were exposed simultaneously to rifampicin and four nucleoside triphosphates [25], is shown in Fig. 1. The RNA-polymerase . dsRNA complexes were extremely sensitive to rifampicin as compared to the case of the enzyme . DNA complexes. This result suggested that the observed low template activity of the dsRNA was mainly due to the slow rate of initiation.

0

10 Rifarnpicin concentration (FM)

20

Fig.1. Effect of rifampicin concentration on the initiation of CPV dsRNA-dependent and calf thymus DNA-dependent RNA synthesis. The standard reaction mixture lacking four nucleoside triphosphates (0.17 ml) was preincubated at 37 "C for 10 min. A solution (0.03 ml) containing four nucleoside triphosphates (final 0.2 mM) and rifampicin (final concentration as indicated), was added. Incubation at 37 "C was continued for 20 min. RNA polymerase activity was defined as [3H]UMPincorporated in the presence of rifampicin as a percentage of that in its absence [25]

0.2

1

0

5

10 15 Metal ions (rnM)

x)

Fig.2. Effect of M n Z f and Mg2 concentrations on CPV dsRNAdependent 1 3 H ] U M P incorporation. Reaction conditions were as described in Materials and Methods except for the indicated amount of metal ions +

Effects of Metal Ions CPV dsRNA-dependent RNA synthesis required Mn2+ ion as described above. As shown in Fig.2, an optimal Mn2+ concentration was 3 mM using nucleoside triphosphate concentrations of 0.2 mM each. Mg2+ was totally inactive at all the concentrations used. Co2+, Ca2+, Cu2+ and Cd2+ were essentially inactive. DNA-dependent RNA synthesis requires either Mg2+ or Mn2+. Addition of Mg2+ increases the synthetic rate at suboptimal level of Mn2+, thus the combination of Mn2+ and Mg2+ has been frequently used in DNA-dependent reactions. Synthetic polyribonucleotide-directed reactions are effectively sup-

182

Double-Stranded RNA Transcription 0.15

I

I 15

Table 2. Template activity of various nucleic acids Reaction conditions were as described in Materials and Methods except that CPV dsRNA was replaced by 10 pg of the listed template nucleic acid and 5 pg of DNase was added where indicated Template

[3H]UMP incorporated - DNdse

fDNase

nmol Q

0' 0

I

I

2

4

6

' '0

8

u subunit (kg) Fig. 3. Effect of u subunit concentration on CPV dsRNA-dependent [ 3 H ] U M P incorporation. Reaction conditions were as described in Materials and Methods except that RNA polymerase was replaced by 30 pg of core enzyme and indicated amount of u subunit. Reaction mixture for T4 DNA transcription contained 0.04 M Tris-HC1 (pH 7.9), 0.2 M KCI, 8 mM MgCI,, 0.1 mM dithiothreitol, 0.1 mM EDTA, 0.2 mM each of ATP, CTP, GTP and [3H]UTP (5 Ci/mol), 5 pg of T4 DNA, 10 pg of core enzyme and indicated amount of u subunit

CPV dsRNA Rice dwarf virus dsRNA Penicillium chrysogenurn virus dsRNA Reovirus dsRNA POIY(A) E. coli 2 3 6 rRNA Calf thymus DNA

0.138 0.128

0.148 0.101

0.292 0.225 0.345 0.035 2.400

0.247 0.031 0.321 0.042 0.081

Table3. RNuse sensitivity of the [ 3 H ] R N A product before and ufter anneuling with CPV dsRNA Preparation of the [3H]RNA product and conditions for annealing and RNase treatment were as described in Materials and Methods Radioactivity in [3H]RNA Annealing

ported by M n2+, and Mg2+ is active with certain ribopolymers [l - 101. With dsRNA as template, however, addition of Mg2 inhibited the Mn2+-activated RNA synthesis. 10mM Mg2+, an optimal concentration for DNA transcription, almost completely inhibited the dsRNA-dependent reaction in the presence of Mn2+. The reason for this inhibition is not clear at present. +

Effect of

CJ

Subunit

The CJ subunit of E. coli RNA polymerase is necessary for the initiation of RNA synthesis at the proper sites on helical DNA template [17,26- 291. As shown in Fig. 3, CPV dsRNA-dependent RNA synthesis was greatly stimulated upon addition of CJ subunit with core enzyme. Little RNA synthesis was observed with core enzyme alone. A comparable stimulatory effect of the same CJ preparation was observed for T4 DNA transcription (Fig. 3). Therefore it is concluded that CJ subunit was indeed required for the transcription of the dsRNA. Various dsRNAs as Templates In order to see whether the observed template activity was restricted to CPV dsRNA, dsRNAs isolated from other viruses were examined for their template activities (Table2). Both rice dwarf virus and Penicillium dsRNAs were active in directing RNA synthesis. The latter showed the highest activity among dsRNAs tested, even though it was ten times less active than calf thymus DNA. However, dsRNA from

total

RNase-resistant

counts/min (% total) 1981 4540

312 (16) 3568 (79)

reovirus has little template activity in our assay system. An apparent incorporation of [3 HIUMP with reovirus dsRNA preparation was diminished in the presence of DNase, indicating contamination of the DNA in the preparation. This observation is in agreement with that reported by Shatkin [12]. Thus, the template activity was significantly different among dsRNA's used. Analysis of the R N A Products The RNA synthesized on CPV dsRNA by E. coli RNA polymerase was isolated and analyzed. The susceptibility of the [3H]RNA product to RNase was examined (Table3). 84% of the RNA was digested by the RNase treatment under the conditions in which single-stranded RNA was completely hydrolyzed but dsRNA was not. This indicated that most of the synthesized RNA had a single-stranded configuration and suggested that the transcription was mainly asymmetric. A minor RNase-resistant fraction (16 %) might be self-complementary RNA products due to partial symmetric transcription, or RNA regions interacting in some way with the template dsRNA [30]. As shown in Table3, the [3H]RNA product hybridized with CPV RNA. The efficiency of hy-

183

M. Sugiura and K. Miura

A

6oC

B

23s

16s

4s

1

1

1

235

16s

4s

1 1

1

600

. 2 .-c

._c

E

E

400

4K

$ 3

9

1

P8!

I 0

2oc

4

i 10

I

I

20

30

9

I

I

CI

200

3

Fraction number

Fig.4. Sedimentation patterns of the [ 3 H ] R N A product in sucrose gradients before ( A ) and after ( B ) denaturation. Preparation of the [3H]RNA, denaturation and centrifugation procedures were as described in Materials and Methods. Arrows indicate the positions of marker RNA’s

bridization was comparable to that obtained with the RNA synthesized by CPV transcriptase within the virions [23], indicating that the product was a true copy of the template sequences. The size of the [3H]RNA product was analyzed by sedimentation on sucrose gradient with and without denaturation. Without denaturation, one major peak at about 13 S was obtained (Fig.4A). However, two main peaks with sedimentation coefficients of around 10 S and 3 S were observed for the denatured sample (Fig.4B). These results suggested that the synthesized RNA chains attached to the template or that they self-annealed, and that the RNase-resistant fraction mentioned above was responsible for the interaction. Starting nucleotide sequences of the RNA labeled with [p3’P]ATP were analyzed. Following hydrolysis with pancreatic RNase, the resulting (5’-32P)-terminal fragments were resolved by DEAE-Sephadex chromatography. As shown in Table4, the distribution of the starting nucleotide sequences of the RNA product was far from a random one. The pppA-Yp fraction (28 %) was subjected to Dowex(X1) chromatography [31] and found to be a mixture of pppA-Cp (10%) and pppA-Up (18%). The pppA-R-Yp fraction was totally resistant to RNase TI digestion and therefore should be pppA-A-Yp (39 %). These results indicated that majority of the RNA chains made on CPV dsRNA were initiated at a few limited sites on the template RNA duplex. The initiation sites were not at the 5’termini of CPV dsRNA (Am-G-Y-) [36]. DISCUSSION Contradictory results have been reported concerning the template activity of viral dsRNA for

Table4. Recovery of (5’-32P)-lubeledfragments produced by digestion of RNA labeled with [y-32P]ATP with pancreatic RNase RNA labeled with [Y-~’P]ATPwas prepared as described in Materials and Methods. It was digested with pancreatic RNase (10 pg/ml, overnight at 37 “C). The hydrolysate (6465 dis./min) was charged onto a DEAE-Sephadex A25 column (0.6 x 20 cm) and chromatographed with a linear gradient from 0.1 M NaCl to 0.4 M NaCl in 0.02 M Tris-HC1 (pH 7.8) containing 7 M urea [24] Fragments

Recovery

% PPPA-YP PPPA-R-YP PPPA-R-R-YP Z pppA-R-R-R-Yp

28 39 11 22

bacterial DNA-dependent RNA polymerases. Gomatos et al. reported initially that reovirus dsRNA was an effective template in RNA synthesis catalyzed by E. coli RNA polymerase [ll]. However, Shatkin showed that reovirus dsRNA purified by a Cs2S04 density gradient did not serve as template for E. coli enzyme and that the apparent RNA synthesis reported by the previous authors was due to the contaminating DNA in their dsRNA preparation [12]. In preliminary investigation with dsRNA isolated from CPV, rice dwarf virus or avian reovirus as template, significant RNA synthesis was not observed [13]. The results reported here clearly show that E. coli DNA-dependent RNA polymerase can use CPV dsRNA as template, although to a lesser extent than DNA. This reaction requires exclusive use of Mn2+ as metal ions. The hybridization experiment indicates that the RNA product is a faithful copy of the template dsRNA. Furthermore, the dsRNA-directed RNA synthesis is dependent on the presence of B sub-

184

M. Sugiura and K. Miura: Double-Stranded RNA Transcription

unit. The effect of CT on DNA transcription has been shown to determine the specificity of the initiation by increasing the affinity of the polymerase for the promoter sites on the template and by decreasing the random affinity of the core enzyme for DNA [26 - 291. Thus, the starting nucleotide sequences of RNA are unique [24,29]. We have shown here that the majority of the RNA products started with ATP have unique nucleotide sequences at their 5’-termini, namely, pppA-A-Y---, pppA-C--- and pppA-U---. These results indicate that transcription in uitro of the viral dsRNA by E. coli DNA-dependent RNA polymerase holoenzyme is initiated at the specific sites. These sites may resemble the promoter regions on E. coli DNA. The fact that the template activities of dsRNA’s from various viruses are significantly different is, therefore, explicable if the promoter-like sequences on dsRNA are assumed not to have uniform distribution among dsRNA’s. One of the important unresolved problems in the transcription mechanism is how a double-helical DNA can be used as a template for the assembly of a single-stranded RNA. Two models have been proposed for this problem. In one, transcription takes place by Watson-Crick base pairing between the growing RNA chain and the DNA template as a result of a local unwinding of DNA after changing from its usual B form to the A form [30,32]. In another, transcription takes place without unwinding the DNA helix and the RNA chain grows in the wide groove of the DNA by a specific stereochemical interaction between the ribonucleotides and the base pairs of the DNA [33]. Relevant to this, it is of interest that the RNA synthesis is supported by the viral dsRNA which exists in a conformation similar to the A form of DNA [34,35]. dsRNA from diplornavirus may be useful as a model template to examine above possibilities.

5. Straat, P. A. & Ts’o, P. 0. P. (1969) J . Biol. Chem. 244, 62636269. 6. Steck, T. L., Caicuts, M. J. & Wilson, R. G. (1968) J. Biol. Chem. 243,2769 - 2778. 7. Hayes, D. H., Cukier, R. & Gros, F. (1967) Eur. J . Biochem. I , 125-134. 8. Hirschbein, L., Dubert, J.-M. & Babinet, C. (1967) Eur. J . Biochem. I , 135-140. 9. Karstadt, M. & Krakow, J. S. (1970) J . Biol. Chem. 245, 746 - 751. 10. Niyogi, S . K. (1972) J . Mol. Biol. 64, 609-618. 11. Gomatos, P. J., Krung, R. M. & Tamm, I. (1964) J. Mol. Biol. 9, 193-207. 12. Shatkin, A. J. (1965) Proc. Nut1 Acad. Sci. U.S.A. 54, 17211728. 13. Miura, K., Sekiguchi, K., Iida, Y. & Kajiro, Y. (1968) Virus 18,555 - 556. 14. Burgess, R. R. & Travers, A. A. (1971) in Procedures in Nucleic Acid Research, vol. 2 (Cantoni, G. L. & Davies, D. R. eds) pp. 851 -863, Harper & Row, New York. 15. Bautz, E. K. F. & Dunn, J. J. (1971) in Procedures in Nucleic Acid Research, vol. 2 (Cantoni, G. L. & Davies, D. R., eds) pp. 743 - 147, Harper & Row, New York. 16. McConnell, D. J . & Br0nner.J. (1972) Biochemisrr-y 11. 43294336. 17. Burgess, R. R., Travers, A. A,, Dunn, J. J. & Bautz, E. K. F. (1969) Nature (Lond.) 221, 43-46. 18. Miura, K., Fujii, I., Sakaki, T., Fuke, M. & Kawase, S. (1968) J . Virol. 2, 1211 - 1222. 19. Kimura, I. & Kodama, T. (1968) Virus, 18, 530-532. 20 Miura, K., Kimura, I . & Suzuki, N. (1966) Virology, 28, 571 579. 21. Takanami, M. (1967) J . Mol. Biol. 23, 135-148. 22. Thomas, Jr, C. A . & Abelson, J. (1966) in Procedures in Nucleic Acid Research, vol. 1 (Cantoni, G. L. & Davies, D. R. eds) pp. 553 - 561, Harper & Row, New York. 23. Shimotohno, K. & Miura, K. (1973) J . Biochem. (Tokyo) 74, 117 - 125. 24. Sugiura, M., Okamoto, T. & Takanami, M. (1969) J . Mol. Biol. 43, 299 - 315. 25. Mandel, W. F. & Chamberlin, M. J. (1974) J. Biol. Chem. 249,2995 - 3001. 26. Travers, A. A. & Burgess, R. R. (1969) Nature (Lond.) 222, 537 - 540. 27. Darlix, J. L., Sentenac, A,, Ruet, A. & Fromageot, P. (1969) Eur. J. Biochem. II,43-48. 28. Bautz, E. K. F., Bautz, F. A. & Dunn, J. J. (1969) Nature (Lond.) 223, 1022- 1024. 29. Sugiura, M., Okamoto, T. & Takanami, M. (1970) Nature (Lond.) 225, 598 - 600. 30. Hayashi, M. (1965) Proc. Nut1 Acad. Sci. U.S.A. 54, 17361743. 31. Okamoto, T., Sugiura, M. & Takanami, M. (1972) Nut. New Biol. 237, 108- 109. 32. Florentiev, V. L. & Ivanov, V. I. (1970) Nature (Lond.) 228, 519 - 522. 33. Riley, P. A. (1970) Nature (Lond.) 228, 522-525. 34. Sato, T., Kyogoku, Y., Higuchi, S., Mitsui, Y., Iitaka, Y., Tsuboi, M. & Miura, K. (1966) J. Mol. Biol. 16, 180-190. 35. Arnott, S., Wilkins, M. H. F., Fuller, W. & Langridge, R. (1967) J . Mol. Biol. 27, 535 - 548. 36. Miura, K., Watanabe, K. & Sugiura, M. (1974) J . Mol. Biol. 86, 31 -48.

The authors thank Miss Y. Murofushi and Mrs H. Sano for their technical assistance, and Dr F. Koide and Dr K. Yazaki for generous gifts of dsRNAs. This work was supported in part by grants from the Ministry of Education and the Ito Science Foundation.

REFERENCES 1. Nakamoto, T. & Weiss, S. B. (1962) Proc. Nut1 Acad. Sci. U.S.A. 48,880 - 887. 2. Krakow, J. S. & Ochoa, S. (1963) Proc. Natl Acad. Sci. U.S.A. 49,88-94. 3. Fox, C. F., Robinson, W. S., Haselkorn, R. & Weiss, S. B. (1964) J. Biol. Chem. 239, 186-195. 4. Niyogi, S. K. & Stevens, A. (1965) J. Biol. Chem. 240, 25872592.

M. Sugiura and K. Miura, National Institute of Genetics, Mishima, Japan 41 1