THE OCCURRENCE OF NATURAL DNA-RNA COMPLEXES IN E. COLI INFECTED WITH T2* BY s. SPIEGELMAN, BENJAMIN D. HALL, AND R. STORCKt DEPARTMENTS OF M1CRC)BlOLOGY ANI) CHEMISTRY, UNIVERSITY OF ILLINOIS ~mrnunicoLed by 7'. M. Sonneborn, May 31, 1961 In a previous publication, Hall and Spiegelman' showed that RNA-DNA com- plexes were formed when mixtures of single-stranded T2-DNA and purifiedZ "T2- specific RNA" were subjected to slow co~ling.~, * It, would appear from this that the Volkin and Astrachanb observation of a similarity in base composition between T2-RNA and DNA is a reflection of a more profound homology. The fact that hybrid formation was found to be unique to the homologous pair led to the conclu- sion that the nucleotide sequences of T2-RNA and DNA are complementary in the sense of the hydrogen bonding scheme proposed by Watson and Crick.G An obvious implication of these findings is that the normal process of transferring information from DNA to the protein synthesizing machine involves a mechanism whereby single-stranded DNA serves as a template for the polymerization of a com- plementary ribopolynucleotide. If continued formation of complementary RNA is a necessary concomitant, it should be possible to find RNA-DNA hybrids in any cell actively engaged in protein synthesis. The T2-E. coli complex was selected as the most suitable system in the initial search for native hybrid. This choice was dictated by the fact that RNA synthesis in a TZinfected cell is virtually confined to the class which is complementary to the viral DNA. The experimental devices employed were in essence similar to those used in the previous' investigation on artificially formed complexes. These in- volved the use of double labeling (HJ and P3*) and equilibrium centrifugation in CsCl gradients' employing swinging bucket rotors. T2-DNA was labeled with Paz by growth of the virus in a medium containing this isotope. T2-specific RNA was marked with H3 by introducing H3-uridine in the early periods of infection. In this experimental setup, the presence of DNA-RNA hybrids would be signalled by the appearance of coincident peaks of the two isotopes. Further, these peaks should occupy a position in the CsCl gradient differing from those which charac- terize the densities of RNA and double-stranded DNA. It is the primary purpose of the present paper to present the results of such ex- periments. The data demonstrate that native hybrids involving DNA and RNA strands do indeed exist. Grmth of celb and conditions of infectiun: Cells of E. coli B were grown in a synthetic medium (medium C of Roberts et aZ.*). As a routine, cella were grown over- night in a glucose limiting medium. They were then harvested, washed, and resuspended in a fresh glucose medium at an 0.D.W of 0.270. Incubation was resumed and continued until logn- rithmic growth wa8 established and an 0.D.W of between 0.285 and 0.300 attained. The cell SUB- pensions were then centrifuged and concentrated tenfold in cold medium lacking glucose. A chilled concentrated hacteriophage suspension was then added in an amount to yield a multiplicity of between 10-30. The mixture was then held at 4°C for 30 minutes to allow for adsorption and then diluted tenfold into fresh glucose-synthetic medium prewarmed to 30°C. Aeration was imme- diately instituted and this moment was taken as the zern time of the experiment. In all experi- ments reported, cell survivors were less than 0.02%. H*-uridine pubes of infeeted celb: In the usual experiment, approximately 300 ml of in- Materiakr and Melhods.--l. 2. 1135 1136 HIO(!HEMIS'I'HY: SPIEGELBIAN ET AL. PROC. N. A. S. fected cells were pulsed at once. Tritiated uridine (1,600 pc-530 pc per pM) at final concentrations ranging from 0.3 pg-1.0 pg per ml was introduced at the appropriate time and the pulse continued for 3 minutes with vigorous aeration. At the end of this period, the entire suRpension was poured on 120 grams of frozen Tris (0.01 M) + NaCl(0.005 M) to stop further incorporation. The cells were harvested by centrifugation, resuspended in 8 ml of the Tris-NaC1 buffer, distributed in separate tubes, and frozen to await centrifugal analysis. 3. Preparation of nucleic acid: The following procedure was finally adopted for `preparing material suitable for the detection of hybrid material in CsCl density gradients. A typical protocol is given. 2.0 ml of frozen cells (equivalent to 80 ml of the original pulsed suspension of infected cells) are thawed, 0.5 ml of 30% sodium dodecyl sulfate added, and the mixture shaken at room temperature for 10 minutes. 0.5 ml of NaC104 are then added and the shaking continued for 5 minutes. This is followed by the addition of 3 ml of chloroform and 0.3 ml of isoamyl alcohol. The mixture is then shaken for 15 minutes following which it is centrifuged in the Servall swinging bucket rotor (HS) for 20 minutes at 6,000. The top aqueous layer is removed and saved. To the residue are added 2.0 ml of Tris-saline and 0.5 ml of 5 M NaC104. The resulting mixture is then shaken for 10 minutes, centrifuged and the aqueous layer removed as previously described. The two aqueous layers are combined and again subjected to the deproteiniiation procedure by adding 6 ml of CHCla and 1.2 ml of isoamyl alcohol. The mixture is then shaken, centrifuged again in the swinging bucket rotor (6K15') and the aqueous layer removed. Two volumes of cold abso`ute EtOH are then added and the mixture held at 0°C for 10 minutes following which the nucleic acid precipitate is collected by centrifugation at 7,000 rpm for 10 minutes. The resuspended pellet is redissolved in Tris (0.001 M) and saline (0.005 M). After solution is complete, the buffer con- centration is brought up to standard strength (0.15 M NaCl, 0.015 M Tris). Cells growing logarithmically in glucose-synthetic, low- phosphate medium (0.002 M Po4 and 0.02 M Tris; pH 7.4) were infected with T2 (multiplicity of infection, 0.02) when an O.Dm of 0.150 was attained. Simultaneously, 10 mc of Pa* were added per liter. The resulting suspension was incubated overnight with shaking at 37°C. After removal of cell debris from the lysate, the virus particles were purified by three cydes of low (3.5K30) and high (12K45) centrifugations. The preparation had 5 x lo-' cpm/particle at the outset. 5. Separation of nucleic acid components by density-gradient centrifugation: The method of fractionation by CsCl density centrifugation is very similar to that described previously.' A layering technique was, however, introduced to cut the centrifugation time required to obtain adequate separation of free RNA from hybrid material. To 0.5 ml of the nucleic acid prepara- tions described in paragraph 3 were added 0.8 gm of solid CsC1. The resulting sdution waa introduced as a bottom layer in a centrifuge tube containing a CsCl solution of density 1.770. This was accomplished by allowing the sample to run through a "disp@pipette" to the bottom of the tube. Prior to its removal, the top of the pipette was plugged with a small stopper to eject the last bit of sample and prevent contamination of the upper portion of the tube when the pipette was withdrawn. The tubes were then placed in the SW-39 rotor of the Spinco model L and centrifuged at 33,000 rpm for 60 hours at 25°C. At the end of each run, fractions corresponding to various density levels in the tube were obtained by piercing the bottom of the tube and collectr ing drops, ca. 30 for each fraction. These were diluted to a volume of 1.2 cc for measurement of ultraviolet absorption and radioisotope concentration. Counting of the two isotopes was carried out as described previously` using trichloracetic acid washing on millipore membranes and the Packard Tri-Carb Scintillation Counter which permits counting of Pa* and HJ in the same sample. Results.-A variety of procedures was surveyed for obtaining material suitable for the reliable detection of DNA-RNA hybrids by CsCl density equilibrium cen- trifugations. These included methods (e.g. freezing and thawing in the presence of IysozymeQ~ 10) of cell rupture and fractionation which conserve the DNA and per- mit its isolation as a relatively intact fraction. However, none of these was re- producibly successful. Ultimately, attempts at preliminary fractionation of cell components were abandoned. Attention was focused on the use of detergents which lysed cells and simultaneously inactivated the enzyme activity of the lysate. It 4. Preparation of PWabeled T2: 6. Counting of Ha-RNA and Paz-DNA: VOL. 47, 1961 BIOCHEMISTRY: SPIEGELMAN El' AL. 1137 was empirically established that removal of most of the protein was a necessary step prior to introducing the material into the CsCl. In the absence of this pre- liminary purification, much of the nucleic acid, including hybridized material, was trapped in the protein layer found floating at the top of the gradient. The procedure described in Methods (section 3) was found to yield highly re- producible results. It consists essentially of the first few deproteinization stages employed for preparing" DNA from bacterial cells. Two modifications were made. None of the steps designed to remove RNA is included, and at the alcohol precipita- tion stage, centrifugation, rather than winding around a glass rod, is used to collect the nucleic acid. Such DNA preparations are, of course, heavily contaminated . with RNA (50% and more) but served well the purposes of the present experiment. The conditions used in the current investigation (synthetic medium, 30°C) were such as to yield a latent period of 3540 minutes and an onset of DNA synthesis sometime after 15 minutes. It is known12- l3 that some uridine ends up in the DNA once its formation begins in the T2-coli B complex. To avoid the complications which this would introduce in identifying radioactive peaks, experiments designed to detect hybrid formation were confined to the period (2-5 minutes) when no DNA synthesis can be detected. In experiments which employ Pa2-labeled T2-DNA and tritiated uridine as the RNA label, existence of hybrid would be detected as a tritium peak in the DNA region, somewhat heavier than the P32 peak corresponding to the T2-DNA. The P32-peak should exhibit signs of bi-modality as evidence of a distribution of the T2-DNA between hybrid and unhybridized T2-DNA. Several independent experiments were performed with essentially similar results. Figures 1 and 2 describe the results of two typical profiles obtained by subjecting nucleic acid preparations from different pulses to equilibrium gradient centrifuga- tion. The general conditions of infection, pulsing, preparation of nucleic acid and separation in the CsCl gradients are all as described under Methods. Detailed differences are found in the legends. The lightest one corresponds to the input P32-labeled DNA of T2. The optical density peak identifies the position of double-stranded E. coli DNA put in as a marker. The tritium peak locates the hybrid. It will be noted that in each instance one observes the existence of a shoulder on the heavy side of the P32 profile correspond- ing in position to the tritium peak region. The relative positions of the peaks and the bi-modality of the P32 distribution are all consistent with what would be pre- dicted from the existence of a DNA-RNA hybrid. If we assume that a small amount of TZDNA can be synthesized even in this early period of infection, another interpretation of the optical density and radio- active profiles can be entertained. This would presume the appearance of new =-DNA, all in the form of single-stranded material, along with the conversion of a portion of the input Ps2-labeled DNA to single strands. To test for this pos- sibility, the alkali lability of the H3 and P32 in relevant fractions of a number of experiments was examined. The same results were obtained in all cases and are exemplified by the data summarized in Table 1. Here, the two fractions (7 and 26) indicated by arrows in Figure 2 were subjected to alkali digestion and the resultant effect on the acid precipitability of H3 and P32 examined. Fraction 7 was included as a free RNA control. It is evident from the data in Table 1 that all of the H3 Three peaks are readily discernible. 1138 BIOCHEMISTRY: SPIEGELMAN ET AL. PROC. N. A. S. f3-5 min. after int) CsCt SWB 33KGOhrs. !b !: 19 -! -00 0. D. ! I \ i ' n3 .40- 9 .20 n 20,000 - 0 0 20 31) 4.0 5.0 ml- FIG. 1.-Adsorption of phage was carried out at a multiplicity of 45. Pa*-labeled phage Pulsing was done between 3-5 minutes after 0 time All other Note that the radioactivity scale is expanded in diluted 2-fold with unlabeled virus. with Ha-uridine having a specific activity of 1,600 pc/pM at a level of 0.3 pg/ml. details are as described under Methods. the hybrid region. TABLE 1 EFFECT OF ALKALI DIGESTION ON ACID PRECIPITABILITY Fraction Acid Prwipitable Counts/ml No. Isotope Control After alkali treatment H3 230 5 26 P3* 735 720 H3 1950 10 7 P32 ... ... 0.4 ml aliquots of Fractions 7 and 26, indicated by arrown in Fin. 2, were made 0.3 N with respect to NaOH and incubated for 24 hours at 30-C. Equivalent aliquots were held an controls under the eame conditions. Following the incubati-n. the alkali warn neu- tralized, carrier herring sperm DNA added, and the contents precipitated with TCA; washing and counting was then carried out as described under Melhoda. and none of the counts are alkali labile. , These data appear to eliminate the possibility that the displaced trit,ium peak can be ascribed to newly synthesized single-stranded DNA. It is also evident that the displ+-.FJ2 cannot be ascribed t,o a conversion of the input viral DNA to an RNA polynucleotide. We may, there- BIOCHEMISTRY: SPIEGELMAN ET AL. - - - - - - P32 500- loo- 00- 1 10 1139 14,000 l0,OOo 6,000 2,000 -- -- H3 300 200 100 VOL. 47,1961 I .6 1.2 0.q .60 0. D. .40 .20 .OO €XI? N-I 7(l) e (PJ2/ + COLI B PT H 3- URIDINE PULSE ! (3-5 min. after inf) ;! CsCl SWB 33K60hrs. II %. i :, h 0 IO 2.0 3.0 4.0 5.0 ml - FIG. 2.-Adaorption of phage was camed out at a multiplicity of 15 of undiluted Pat labeled virus. pulsing was done between 3-5 minutes after 0 time with Ha-uridine having a specific activity of 530 pc/& at a level of 1 fi /ml. AU other details are as described un- der Methods. Note that the radioactivity scafe is expanded in the hybrid region. fore, conclude that a hybrid region exists containing newly synthesized T2-specific RNA complexed with some of the input Pa2-labeled DNA. Discussion.-The existence of native DNA-RNA hybrids in "2-infected E. coli is consistent with the assumption that DNA serves as a template for the synthesis of complementary informational RNA. This view would require the existence of a DNA-dependent enzymatic mechanism for synthesizing polyribonucleotide. Evidence suggesting the presence of such a pathway emerged from a study" of cell-free preparatiowderived from E. coli which possessed considerable capacity to synthesize polyribmdieotide. The observed synthetic activity exhibited a re- quirement for riboside tri-phosphates and Mn, and was sensitive to DNAase. More recently a $umber of laboratorie~~~-~~~ 2o have independently achieved con- siderable purific&ion of an RNA polymerase capable of synthesizing polyribo- nucleotide from gbose triphosphate in the presence of DNA. Furthermore, the base ratio of the Polyribonucleotide synthesized bears a striking homology to that of the DNA udas a l9 It would appear from the preliminary reports available that considerable information will soon be available on the enzymological 1140 BIOCHEMISTRY: SPIEGELMAN ET AL. PROC. N. A. S. aspects of DNA-mediated RNA synthesis. It will be of obvious interest to see whether hybrid formation is the first step in this process. The experiments described in the present paper were designed primarily to detect the presence of native RNA-DNA hybrids. The use of isotopic labeling combined with CsCl gradients in swinging buckets provides a sensitive method for the certain detection of hybrid even when it is present in minute quantities. However, these methods are too laborious to permit the rapid accumulation of the kind of detailed quantitative and kinetic information we would like to possess. Other more efficient and less time-consuming methods for both the detection and estimation of hybrid material will undoubtedly be developed as we learn more of its nature. It may per- haps be of value for further efforts if we summarize briefly here the experience we have thus far accumulated with the T2-E. coli complex. As determined by their relative position, the densities of the natural hybrids are very similar to those formed artificially by the slow cooling process. In seven experiments involving pulses between 3 and 5 minutes after infection, the proportion of the newly synthesized RNA found in the hybrid region ranged between 1 and 0.1 per cent. As measured by the amount of Paz shifted into the hybrid region, the proportion of input DNA hybridized varied from 15-30 per cent in all cases where approximations could be made. However, it may be noted that hybrids involving the input DNA were observed in pulses covering both 9-12 minutes and 19-22 minutes after infection. Summary.-Experiments are described exhibiting evidence for the existence of natural DNA-RNA hybrids in E. coli cells infected with T2. The use of double labeling (PS2-DNA and H3-RNA) coupled with equilibrium centrifugation in CsCl gradients in swinging bucket rotors permitted the isolation and identification of the hybrid regions. ance of the experiments described here. colleagues, Professors K. C. Atwood and N. Sueoka, are gratefully acknowledged. Our experience with other stages of infection are not as extensive. The authors would like to thank Mr. Saul Yankofsky for his skillful assistance in the perform- In addition, the many helpful discussions with their * This investigation was aided by grants in aid from the U.S. Public Health Service, National t Present address: Bacteriology Department, University of Texas, Austin, Texas. 1 Hall, B. D., and S. Spiegelman, these PROCEEDINGS, 47, 137 (1961). * Nomura, M., B. D. Hall, and S. Spiegelman, J. Molec. Bwl., 2, 306 (1960). * Marmur, J., and D. Lane, these PROCEEDINGS, 46, 453 (1960). 4 Doty, P., J. Marmur, J. Eigner, and C. Schildkraut, op. cit., 46, 461 (1960). 6Volkin, E., and L. Astrachan, Virology, 2, 149 (1956). 6 Watson, J. D., and F. H. C. Crick, Nature, 171, 964 (1953). 7 Meselson, M., F. Stahl, and J. Vinograd, these PROCEEDINGS, 43, 581 (1957). 8 Roberts, R. B., P. H. Abelson, D. B. Cowie, E. T. Bolton, and R. J. Britten, Studies of synthesis in E. coli (Washington: Carnegie Institution of Washington, 19571, p. 5. * Kob, A., J. Bacteriol., 79, 697 (1960). 11 Marmur, J., J. Molec. Biol. (in press). I* Cohen, S. S., H. D. Barner, and J. Lichtenstein, J. Bwl. Chem. (in press). I* Ben-Porat, T., "RNA Metsbolism in Phage Infected Protoplasts," Doctoral Thesip, Uni- 14Spiegelman, S., Recent Progr. Microbiol., 7, 82 (1959). 16 Weiss, S. H., these PROCEEDINGS, 46, 1020 (1960). Science Foundation, and the Office of Naval Research. Hall, 8. D., R. Storck, and S. Spiegelman, J. Biophys. (in press). versity of Illinois (1959). Vor,. 47, 1961 BIOCHEMISTRY: N. SIJEOKA 1141 l6 Hurwitz, J., A. Breuler, and R. Dinger, Biochettl. Biophys. Research Contn~un., 3, 15 (1960). 18 Furth, J. J., J. Hurwitz, and M. Goldmann. op. cit., 4, 362 (1961). le Weiss, S. B., and T. Nakamoto, these PROCEEDINGS, 47, 694 (1961). a Ochoa, S., D. P. Burma, H. Kroger, nnd J. D. Weill, op. cit., 47, 670 (1961). Stevens, A., op. cit., 3, 92 (1960).