Reprinted from the Proceedings of the NATIONAL Aca~Exu OF SCIENCES Vol. 47, No. 2, pp. 137-146. February, 1961. SEQUENCE COMPLEMENTARITY OF 7'2-DNA AND TL-SPECIFI(' RNA* BY BENJAMIN D. HALL AND S. SPIEGELMAN DEPARTMENTS OF CHEMISTRY AND MICROBIOLOGY, UNIVERSITY OF ILLINOIS Communicated by H. E. Carter, December 8, 1960 Investigations of the functional interrelations among DNA, RNA, and protein are most conveniently performed under conditions which limit the synthesis of each macromolecular class to a few chemical species. A situation of this type obtains in E. coli cells infected with bacteriophage T2. Volkin and Astrachan' examined the nature of the RNA synthesized in the T2-coli complex by means of 1'32-labeling. Estimation of the relative P32 content of the 2',3'-nucleotides isolated from an alkaline hydrolysate led Volkin and Astrachan to deduce that the RNA formed in the infected cell possessed an apparent base ratio analogous to that of T2-DrU'A. Subsequently, Volkin2 obtained data suggesting that the synthesis of a specific RNA is a prerequisite for the intracellular production of bacteriophage. Nomura, Hall, and Spiegelman3 confirmed the observations on the apparent base ratios. In addition, they offered independent evidence for the existence of a "T2-Specific RNA" by demonstrating that RNA molecules synthesized after in- fection differed from the bulk of the E. coli RNA in electrophoretic mobility and average sedimentation coefficient. Because the procedures employed (zone electro- phoresis and sedimentation) led to a selective separation of T2-specific RNA from the normal RNA of E. coli, they open up possibilities of further experiments relevant to an understanding of the nature of T2-RNA. The fact that "T2-RNA" possesses a base ratio analogous to that of T2-DNA is of interest because it suggests that the similarity may go further and extend to a detailed correspondence of base sequence. The central issue of the significance and meaning of "T2-RNA" is whether or not this is in fact the case. X direct attack on this problem by complete sequence determination is technically not feasible at the moment. However, some recent findings of Marmur, Doty, et ~1.~3~ suggest the possibility for an illuminating experiment. These authors demonstrated the specific reformation of active double-stranded DNA when heat-denatured DNA is subjected to a slow-cooling process. Such reconstitution of the double- stranded structure occurs only between DNA strands which originate from the same or closely related organisms. Presumably, the specificity requirement for a successful union of two strands reflects the need for a perfect, or near-perfect, complementarity of their nucleotide sequences. We have here then a possible method for detecting complementary nucleotide sequences. The formation of a double-stranded hybrid during a slow cooling of a mixture of two types of poly- nucleotide strands can be accepted as evidence for complementarity of the input strands. 137 138 BIOCHEMISTRY: HALL AND SPIEGELMAN PROC. N. A. S. We have used this procedure to examine for complementarity of sequence between "T2-RPII'A'' and T2-DNA. Purified T2-RNA was used in order to provide an optimal opportunity for the T2-RNA to combine with its DNA complement, un- hindered by non-specific interactions involving irrelevant RNA. Since the hybrid would have a lower density than uncombined RNA, a separation of the two might be attainable by equilibrium centrifugation in CsCl gradients.6 To insure a sensitive and unambiguous detection of the hybrid, should it occur, double labeling was used. The T2-RNA was marked with and the T2-DNA with H3. Two isotopes emitting &particles differing in their energies are conveniently assayed in each other's presence in a scintillation spectrometer.' This device, coupled with the use of the swinging-bucket rotor for the equilibrium centrifugation, permits the actual isolation of the pertinent fractions along with a ready and certain identifica- tion of any hybrids formed. The primary purpose of the present paper is to present the results of such experi- ments. The data obtained demonstrate that specific complexes are indeed formed between "T2-RNA" and its homologous DNA. Their occurrence offers strong pre- sumptive evidence for a detailed complementarity of the nucleotide sequences in these two macromolecules. 1. Preparation and denaturation of DNA: Tritiated phage were prepared by the addition of Ha-thymidine to a T2-infected culture of E. cola B. The cells were treated with 5-fluorouracil deoxyriboside (0.5 pg/ml) prior to infecton. The phage were purified by treatment of the lysate with DNAase and RNAase followed by three cycles of high- and low-speed centrifugation. DNA was extracted from the purified phage by treatment with sodium dodecyl sulfate followed by chloroform-iso-amyl alcohol deproteinization and ethanol precipitation of the DNA.8 This preparation will be designated by H3-DNA(T2). DNA from other sources was similarly purified. Tritiated E. cola DNA was prepared from cells of a thymineless mutant (15T-) grown in a synthetic medium supplemented with tritiated thymidine. This preparation will be designated by H3-DNA (E. coli). DXA used for complex formation with RNA was first denatured by heating for 15 minutes at 95°C in 0.15 M NaCl + 0.01 M sodium citrate (pH 7.8), after which the tube containing the DNA was quickly placed in an ice bath. In all cases, the denaturation was carried out at a DNA concentration of 130 pg/cc. P32-labeled ribosome RNA was obtained from E. Cola B grown in synthetic medium, infected with T2 at a multiplicity of 3.8, and labeled with 10 millicuries of The infection and radioisotope incorporation were done at 37°C in medium C (Roberts et ~1.)~ modified to include 5 gm NaC1, 0.37 gm KC1, and 1 gm casamino acids per liter. The phosphate concentration was lowered to M and 0.1 M tris (hydroxymethyl) aminomethane (tris), pH 7.3, was used for buffering. The number of infective centers and uninfected survivors (2.5%) agreed with the multi- plicity of infection. The procedures used for stopping incorporation, washing and disrupting cells, and preparing ribosomal RNA were those described previously (Nomura, Hall, and Spiegel- man3). 3. Purz$cation of T2-specijic RNA: Enrichment of the ribosomal RNA preparation in its content of TZspecific RNA (as judged by an eightfold increase in specific activity of P32) was obtained by zone centrifugation through a sucrose gradient. One ml of a 1.5% sucrose solution (w/w) + one ml of P32 ribosome RNA solution (1 mg/ml, 106,000 cpm/ml) were layered, with an inverted gradient of RNA on 20 ml of a 2 to 15 per cent sucrose gradient. All solutions were 0.05 M in KC1 and 10-2 M in tris buffer at a pH of 7.3. Following centrifugation for eight hours at 25,000 rpm in the SW-25 rotor of the Spinco preparative ultracentrifuge, the contents of the tube were removed by dripping through a hole punctured in the bottom of the tube. Fractions of 1.2 ml were collected by drop counting. The ultraviolet absorption at 260 mp and P32 content of the fractions are shown in Figure 1. The two fractions at the peak of P32 activity (corresponding 2. Preparation of T2-speczjic RNA labeled wzth P32: between three and eight minutes after infection. to 18 and 19.8 1111) were used for hybrid formation. RNA(T2). 0.03 ilf in sodium citrate and 0.3 112 in NaCl at a pH of 7.8. capacity of 40 liters was used to provide slow cooling as follows: These will be referred to in the text as In the experiments described below, the two fractions exhibited identical properties. Slow cooling of DNA and RNrl: Slow cooling of RNA and DNA was done in solutions An insulated water bath having a 4. Time (hr.) 0 3 . 5 7.5 13.5 24.5 30.5 Temperature ("C) 65 52 44 3 0 28.5 26 When the bath temperature reached 26", the tubes curitairling RNA arid I)NA were removed aid brought to a volume of 5.1 cc and a density of 1.74 gm/cc by addition of suitable amounts of water and saturated CsCl solution. Twenty-five Gg of unlabeled, undenatured DNA were added to the solution as a reference density marker. Separation of RNA from DNA by density-gradient centrifugation: The solutions of RNA and 5. DNA containing CsCl were centrifuged at 33,000 rpm in the SW-39 rotor at a temper- ature of 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 collecting drops, 30 for each fraction. These were diluted to a volume of 1.2 cc for measurement of ultraviolet absorption and radioisotope concentration. Counting of HiDNA and P32-DNA: To an aliquot from each swinging-bucket fraction 250 pg herring sperm DNA was added as carrier. The nucleic acid was then precipitated with trichloracetic acid (final concentration 107,) in the cold, collected, and washed on a millipore filter (course, 50 mm dia.). The filter was air- dried for one hour and placed in a cylindri- cal glass vial filled with 15 ml redistilled toluene containing 1.5 mg of 1,4bis-2- (5-phenyloxazolyl)benzene (POPOP) and 60 mg of 2,5-diphenyloxazole (PPO). P32 and H3 were counted in a Packard Tri- Carb liquid scintillation counter. 6. 4.0 0 lo RI 0 20 0 0 SWB [ 25 K 4801 OF RIBOSOME - RNA OF Tz-INF. E. COLI. .... .. .. : . .. .. .. I n 0 000 500 I I' 0 ,...... . 5 10 15 20 25 - ml- FIG. 1.Separation of P32-RNA(T2) from E. Coli RNA by sucrose-gradient sedimentation. One ml. of ribosome RNA solution containing 1 mg RNA and 106,000 cpm P32 + 1 ml 1.5% sucrose solution were layered on 20 ml of a 2-15% sucrose gradient. Centrifugation: 8 hours at 25,000 rpm. Cpm shown refer to 0.05 ml fractions of the swinging- bucket fract,ions. Separation of T2-Xpecific RNA in a CsCZ Gradient.-It was first necessary to establish the conditions required for an adequate separation of T2-specific RNA from TB-DNA. Whereas E. coli ribosome RNA formed a narrow band within two days, T-2 RNA, because of its smaller size, required five days to form a band near the bottom of the tube. Figure 2 shows the result of a five-day run carried out under the conditions speci- fied above. The mixture being separated consisted of 6.5 pg of heat-denatured H3-DNA(TB), 25 pg of unlabeled and undenatured T2-DNA, and 14 pg of the purified P32-RNA(T2). Here, the three nucleic acids were not exposed to a slow- cooling operation but were mixed at 25"C, immediately put in the CsCl solution, 140 BIOCHEMIS1'RY: HALL AND SPIEGELMAN PROC. N. A. S. and centrifuged. It will be noted that there is no appreciable irit,eraction between the RNA and DNA as evidenced by the absence of any appreciable overlapping of the and H3-containing regions. The small "tail" of which extends to the top of the tube is presumably a consequence of the low molecular weight of the T2-specific RNA. Hybrid Formation between Denatured T2-DNA and T2-Specific RNA .-The re- sults described in Figure 2 show that CsCl density gradient centrifugation permits a clear separation of H3-DNA(T2) from P32-RNA(T2) and provides, therefore, a test for interactions leading to the formation of RNA-DNA hybrids. Any distor- tion of the distribution of H3-DNA or P32-RNA from that observed in Figure 2 which leads to regions of overlap between H3 and P32 would be indicative of such interac- tions. .20 0 u) N e 0 .IO 0 0 5 10 15 20 25 30 35 FRACTION NUMBER FIG. 2.-Separation of PaZ-RNA(T2) from H3-DNA(T2) by CsCl- gradient centrifugation. A mixture of 6.5 pg heat-denatured Ha- DNA, 14 pg PS2-RNA and 25 pg undenatured, unlabeled T2 DNA was made at. 25"C, immediately diluted with CsCI, and then centri- fuged for five days at 33,000 rpm. 1. The eflect of temperature during slow cooling on hybrid formation: The influence of the starting temperature of the slow-cooling process was examined in a number of runs. In all cases, the nucleic acid mixture incubated consisted of 6.5 pg of heat-denatured HS-DNA(T2) and 14 pg of P32-RNA(T2). The rate and conditions of the cooling were as described earlier. Three tubes containing this RNA-DNA mixture were placed in the slow-cooling bath at starting temperatures of 65", 52", and 40°C respectively. Slow c-ooling was followed by CsCl gradient centrifugation. obtained at the three temperatures. Comparison of the profiles of H3 and P32 with those of the control (Fig. 2) shows clearly that in all three cases, slow cooling of the DNA and RNA has produced a new peak of approximately centered on the band of H3 (denatured DNA). This new P2-containing band must contain an RNA-DNA hybrid having approximately the same density as denatured T2-DNA. The amount of complex formed on cooling from the three temperatures was the Figure 3 shows the optical density profiles and distributions of H3 arid VOL. 47. 1961 BIOCHEMISTRY: HALL AND SPIEGELMAN 141 0 5 IO 15 20 25 30 35' - 200 - 0 5 IO 15 20 25 30 35 FRACTION NUMBER FIG. 3.-Formation of DNA-RNA hybrid at various temperatures. CsC1-gradient centrifugation analysis. P32-RNA(T2) ( 14 pg) and H3-DNA(T2) (6.5 pg) were mixed in 0.6 ml 0.3 M NaCl and 0.03 M Na citrate, (pH 7.8); then the solution was immediately placed in the slow-cooling bath. Three identical solutions were made; (a) was placed in the bath at 65", (b) at 52", and (c) at 40°C. When the bath temperature reached 26", CsCl and 25 pg T2 DNA were added to each solution; then they were centrifuged for five days at 33,000 rpm. 142 BIOCHEMfSTRY: HALL AND SPIEGELMAN PROG. N. A. 9. same within experimental error. The thrce differ slightly in the density of the complex relative to DNA, the complex formed at low temperature being apparently more dense. This may be explained by the occurrence of partial renaturation of the H3-DNA at the higher temperatures. Requirement for presence of single-stranded DNA during cooling: In order to successfully complex with TZ-RNA, the molecules of TZ-DNA must be present, in the single-stranded state. This was shown by an experiment in which a mixture of native H3-DNA(T2) (13 pg) and P2-RNA(T2) (15pg) was subjected to slow cool- ing, starting from 40°C. No evidence of hybrid formation is observed (Fig. 4). In a companion run (a repetition of the experiment of Fig. 3c) with denatured H3- DNA(T2), approximately 10 per cent of the P3"RNA was included in the hybrid region. 2. 0 ul *. d 0 0 5 IO 15 20 25 30 FRACTION NUMBER FIG. 4.--CsCLgradient centrifugation of a slowly cooled mixture of native H3 DNA (T2) with P32-RNA(T2). 13 pg QNA and 15 pg RNA were mixed in 1.2 ml 0.3 M NaCl. 0.03 M Na citrate, slowly cooled from 40" to 26", diluted with saturated CsCl solution, and centrifuged. 3. Stoichiometry of hybrid formation: Assuming the specific activity of 1'32 to be equal in all RNA molecules, one can estimate the amount of RKA which formed hybrid. (This figure is a maximum value, for some of the RNA which failed to form hybrid may be pre- existing E. coli RNA and, therefore, devoid of F.) The amount of DXA in the hybrid cannot exceed 6.5 pg, the total amount of denatured TZ-DKA present. Be- cause the hybrid and denatured T2-DNA have the same density, no more precise estimate can be made. From these considerations, it appears probable that the ratio of DNA to RNA in the hybrid does not exceed 5. That the complex does in fact contain considerably more DNA than RNA is suggested by its density, which is very nearly that of T2-DNA. A further indication that the entire 6.5 pg of TBDNA has participated in complex formation is the proportionality observed between the amount of hybrid formed and the amount of DNA present when the rat)io of DNA to RNA is varied. Experiments completely comparable to those From the data of Figure 3, this would be 1.3 pg RNA. VOL. 47, 1961 BIOCHEMISTRY: HALL AND SPIEGELMAN 143 described by Figure 3 werc carried out with the same wnwntration of P2-RNA(T2) but with In these cases, the amount of P2-RNA found associated with the denatured DNA band was approximately two per cent of the input RNA which compares with the average of ten per cent ob- served when five times as much T2-DNA is included in the cooling mixture. On the speciJicity of the interaction leading to hybrid formation: Having es- tablished the existence of the phenomenon and the conditions required for its oc- currence, it became of obvious interest to examine the specificity requirements of hybrid formation. This was tested by carrying out the cooling process with mix- tures of P32-RNA(T2) and denatured DNA from heterologous sources. These in- cluded DNA of Ps. aeruginosa, E. coli, and bacteriophage T5. The DNA of the E. colz was labeled with H3-thymidine whereas the others were unlabeled. The mixtures of P32-RNA(T2) and denatured DNA preparations were subjected to a 52°C slow-cooling incubation under conditions identical to those described for the experiment of Figure 3b. Upon completion of the incubation, unlabeled native DNA was added to each tube as a density marker. Figure 5 gives the optical density profiles and distributions of radioactivities. There is a suggestion of a very slight peak of P2 in the DNA region of the mixture containing E. coli DNA. It coresponds to l/30 the amount of hybrid produced with T2-DNA in a similar experiment. This may reflect the presence of the small amount of non-infected cells present during the P32-labeling of the material from which the T2-RNA was obtained. However, it is too small to be considered seri- ously without further investigation. None of the other heterologous mixtures tested yielded detectable amounts of hybrid. It is of interest to note that although T5 has the same over-all base ratio as T2, no evidence of interaction between T2-RNA and T5-DNA was observed. Interpretation of the Results.-The data presented here show that RNA molecules synthesized in bacteriophage-infected cells have the ability to form a well-defined complex with denatured DNA of the virus. That this interaction is unique to the homologous pair is shown by the virtual absence of such complexes when T2-specific RNA is slowly cooled with heterologous DNA. The fact that T2-RNA and DNA do satisfy the specificity requirement must reflect a correspondence in structure between the two. Structural specificity of this order in single polynucleotide strands can only reside in definite sequences of nucleotides. We conclude that the most likely interrelationship of the nucleotide sequences of T2-DNA and RNA is one which is complementary in terms of the scheme of hydrogen bonding proposed by Watson and Crick.Io Extension to Other Systems.-The bulk of the RNA in E. coli corresponds to the 18s and 25SI1 components of the ribosomes. These are metabolically stable,'? remain firmly attached to ribosome protein at M Mg++,13 and have a base composition14 not related in any obvious way to the DNA of the cells. In addition to the lark of correspondence in base ratio, two other reasons can be advanced for doubting the suitability of the large ribosomal RNA molecules for directing the synthesis of proteins as specified by the genetic material. First, the experiments of Riley el. al.I5 suggest that the intermediary between the genome and the protein- synthesizing mechanism is metabolically unstable. Second, the formation of the larger RNA components is virtually absent in T2-infected cells (cf. Fig. 1) despite the quantity of denatured T2-DNA. 4. 144 c2 2c 0 W N D 0 10 0 @ 01 0 W N 0 00 BIOCHEMISTRY: HALL AND SPIEGELMAN PROC. N. A. s. .. _. :: .. .: i. .. .. .. .. .. .. .. I2O0 f PSf- RNA +DNA z a 100 * 0 200 5 a 100 * 0 FIG. 5.4sCl-gradient centrifugation of slowly cooled heterologous DNA-RNA mixtures. (a) P32-RNh(T2) + H3-DNA (E. coli), Ih) P32-RNA(T2) + DNA (Ps aeruginosu), (c) P32-RNA(T2) + DNA nr~%us DNAtO (b). VUL. 47, 1961 HI OC'HEMlS!l'li?Y: HALL AN U SPIhGELM AN 145 the fact that they are actively synthesizing a variety of new protein species. It seems more likely that the RNA molecules directly concerned with specifying protein synthesis in normal cells would have a base ratio corresponding to DNA and would possess other properties analogous to those found for T2-specific RNA. Its principle characteristics may be summarized as follows:', (1) a weak linkage with the ribosome fractions since it can be broken by dialysis against M Mg++, (2) an active metabolic turnover, (3) an average sedimentation coefficient of about 8S, (4) a base composition which is closely analogous to its homologous DNA (considering thymidine equivalent to uridine and similarly for cytidine and hy- droxymethyl cytidine), and (5) a sequence complementary to its homologous DNA. The detection of the complementary RNA in T2-infected cells was greatly facili- tated by the fact that the larger ribosomal components are not synthesized. In- deed, it would appear as if RNA synthesis in the T2-coli complex is largely confined to the class which is complementary to DNA. This advantage is not present in uninfected cells. Consequently the search for normal complementary RNA will be technically more difficult. That it is nevertheless feasible is suggested by the experiments of YEas and Vincent16 with yeast. in a manner comparable to the procedures of Volkin and Astrachan and, despite surprisingly long pulses, they were able to detect the formation of a fraction with a high meta- bolic turnover and possessing a base composition analogous to yeast DNA. Ultimately, attempts at establishing the presence in normal cells of RNA com- plementary to the genetic material will require that it be separated from the other RNA components. If all complementary RNA molecules possess physical chemical characteristics analogous to those of T2-specific RNA, the same methods which effected a successful isolation in this case may well serve in others. Once isolated, sequential complementarity to relevant DNA can be examined by the methods described above. Some Implications of Complementary RNA.--,4n increasing amount of attention is currently being focused on the possibility of forming hybrid helical complexes composed of DNA and RNA strands. Interest in this originates quite naturally from its obvious implication for translating the genetic information coded in DNA to a functional RNA complement. Previous experimentsl7~ l8 had already demon- strated that paired helices were generated in mixtures of the synthetic polyribo- nucleotides of uridylate and adenylate. More recently,lg, *O this has been ex- tended to combinations involving synbhetic polydeoxyribonucleotides and poly- ribonucleotides. It is of some interest that the experiments reported in the present paper lend support to the concepts underlying such model experiments by exhibiting hybrid formation between natural polyriucleotides which are complementary and biologically related. The demonstration of sequence complementarity hetween homologous DNA and RNA is happily coiisistent with an attractively simple mechanism of informational RNA synthesis in which a single strand of DNA acts as a template for the poly- merization of a complementary RNA strand. Summary.-Experiments are described showing specific complex formation between single-stranded T2-DNA and the REA synthesized subsequent to infec- tion of E. coli with bacteriophage T2. No such hybrid formation is observed with These authors used heterologous DNA even if it has the same over-all base composition as TB-DNA. It is concluded that T2-DNA and TBspecific RNA form hybrids because they possess complementary nucleotide sequences. The generality of the existence of complementary RNA and its possible role as a carrier of information from the gen- etic material to the site of protein synthesis is briefly discussed. We are indebted to Drs. Noboru Sueoka and John Drake for providing the T5 DNA and Pseudomonas DNA used in this work and to Miss Cherith Watson for her assistance in performing the experiments. * This investigation was aided by grants in aid from the U.S. Public Health Service, Natioiial Science Foundation, and the Office of Naval Research. * Volkin, E., these PROCEEDINGS, 46, 1336 (1960). 3 Nomura, M., B. D. Hall, and S. Spiegelman, J. Mol. Bid, 2, 306 (1960). Marmur, J., and D. Lane, these PROCEEDINGS, 46,453 (1960). SDoty, P., J. Marmur, J. Eigner, and C. Schildkraut, ibid., 46, 461 (1960). 6 hteselson, M., F. W. Stahl, and J. Vinograd, ibid., 43,581 (1957). 7 Okita, G. T., J. J. Kabara, F. Richardson, and G. V. Le Roy, Nucleonics, 15, 111 (1957). gRoberts, R. B., P. H. Abelson, D. B. Cowie, E. T. Bolton, arid R. J. Britten, "Studics of Biosynthesis in Escherichia coli," Carnegie Institution of Washington, Publication f607 ( 1957), Volkin, E., and L. Astrachan, Virology, 2, 149 (1956). Marmur, J., personal communication. p. 5. lo Watson, J. D., and F. H. C. Crick, Nature, 171, 964 (1953). l1 Kurland, C. G., J. Mol. Biol., 2, 83 (1960). l2 Davern, C. I., and M. Meselson, ibid., 2, 153 (1960). l3 TissiBres, A., J. D. Watson, D. Schlessinger, and B. R. Hollingworth, ibid., 1, 221 (1959). I4 Spahr, P. F., and A. Tisskes, ibid., 1, 237 (1959). Riley, pvl., A. B. Pardee, F. Jacob, and J. Monod, ibid., 2, 216 (1960). YEas M., and W. S. Vincent, these PROCEEDINGS, 46, 804 (1960). Warner, R. C., Fed. Proc., 15, 379 (1956). l8 Rich, A., and D. R. navies, J. Am. Chem. Soc., 78, 3548 (1956). l9 Rich, A., these PROCEEDINGS, 46, 1044 (1960). 2o Schildkraut, C. L., J. Marmur, J. R. Fresco, and P. Doty, J. Biol. Chem. (in press).