THE Jonarur. osBnxoorcn~ CHE~BT~Y Vol.238. No.9,Septembx1963 Printed in U.S.A. Chemically Synthesized Deoxypolynucleotides as Templates for Ribonucleic Acid Polymerase" ART~RO FALASCHI,~ JULIUS ADLER, AKD H. G. KHORAK~ From the Departments of Biochemistry an,d Genetics and the Institute for Enzyme Research., the Unioersity of Wisconsin, Madison 6, Wisconsin (Received for publication, April 25, 1963) The enzymatic synthesis of ribonucleic acid from the four ribonucleoside 5'-triphosphates in the presence of deoxyribo- nucleic acid has been documented in a number of laboratories (1-11). The reaction is catalyzed by an enzyme called RNA polymerase and the DNA determines the composition and the sequence of nucleotides in the synthesized RNA according to the base-pairing principle first recognized for DNA by Watson and Crick (7-11). The RNA synthesized in this manner stimu- lates the incorporation of amino acids into protein (12-15). A deoxypolythymidylate fraction obtained by chemical syn- thesis (16) was tested previously by Hurwitz et al. (7, 11) and was found to bring about the synthesis of ribopolyadenylic acid. In the present work, purified synthetic deoxypolynucleotides of known size and sequence have been used to study further the mechanism of action of RNA polymerase. The results reported herein show that deoxyoligonucleotides as small as the penta- nucleotide dTs can serve as templates for the synthesis of ribo- polyadenylate.' The rate of the synthetic reaction increases with an increase in the size of the deoxypolythymidylate until a chain length of 14 is reached, and this polymer was actually more active than DNA. The product synthesized from deoxy- polythymidylate of various sizes was invariably very much larger than the size of the template. Several lines of evidence showed that the enzyme initiated the synthesis of new chains, rather than causing esterification to the 3'hydroxyl end of the deoxy- polynucleotide chains. Finally, in these simpler systems, the incorporation of the ribonucleoside 5'-triphosphates again fol- lowed the Watson-Crick base-pairing principle, although some exceptions were noted. A brief report. of these findings has already appeared (17). * Supported by grants from the Nntional Institutes of Health, the National Science Foundation, Life Insurance Medical Re- search Fund, and the Graduate School of the University of Wis- consin. t Fellow of the International Laboratory of Genetics and Bio- physics, Naples, Italy. Present address, Department of Bio- chemistry, Stanford University Xedical School, Palo Alto, Gali- fornia. 1 All of the chemically synthesized polynucleotides used in the present work carry a 3'.hydroxyl group at one end and a 5'-phos- phate group at the opposite end. For convenience, this class of homologous polynucleotides is simply designated in the text by the nncieoside initial with a subscript indicating the number of nucleosides in the chain. For example. deoxv-DTn'f'oTnToT is dTI and deoxy-pCpCpCpCpCpC is dCs,`etc. "A, G: c`, U, $nd T stand for adenosine, guanosine, cytidine, uridine, and thymidine. EXPERIMEKTAL PROCEDURE RNA polymerase was purified from Escherichia coli according to the procedure described by Chamberlin and Berg (10). E. coli phosphodiesterase (18) was a gift from Dr. I. R. Lehman. C"- Ribonucleoside 5'-triphosphates were purchased from Schwarz BioResearch, Inc. Deoxytibopolynucleotides-Deoxycytidine, deoxyadenosine, and deoxyguanosine oligonucleotides were prepared by published procedures (19-21). The homologous deoxythymidine poly- nucleotides dT1 to dTll were prepared as described previously (16, 22), and C?d-labeled thymidine polynucleotides were pre- pared by Mr. W. J. Connors by adaptation (23) of these pub- lished procedures (16, 22). Pure thymidine polynucleotides dTIP to dTr4 and a fraction containing members higher than dTr4 were prepared from the 1 M triethylammonium bicarbonate fraction obtained in an experi- ment described previously (Khorana and Vizsolyi, Table I, and Fig. I (16)). The latter fraction, as mentioned earlier (16), was a complex mixture of polynucleotides containing apparently a high proportion of oligonucleotides linked together by pyrophos- phate bridges between the terminal 5'-phosphomonoester groups, n'ith general structure PT(PT),,PT 0' \ PT(PT)~PT The mixture (110 optical density units at 267 rnp) was treated in dry pyridine (2 ml) with 0.5 ml of acetic anhydride for 3 days in order to selectively cleave the pyrophosphate bonds, and the reaction mixture was worked up as described earlier (22). Chro- matography on a DEAE-cellulose (carbonate form) column (50 X 1.2 cm) gave a series of peaks which were processed by the standard method developed earlier (16). As a result of cleavage of the pyrophosphate bonds, more than 50% of the ultraviolet- absorbing material was present as a mixture of oligonucleotides smaller than dTro. The higher homologues were then applied to paper along with previously characterized dT,, as marker. The dTr*, dT13, and dTld traveled with progressively decreasing RF values when chromatographed for 2 weeks in descending n-propyl alcohol-concentrated ammonia-water (55 : 10: 35). Acetylation of J'-Hyd~o~yl End Groups in Penta- and Undeca- thym.idylic Acids--An aqueous solution of ammonium salt of the polynucleotide (2 pmoles of thymidine) was passed through a column (1 x 2 cm) of Dowex 50 ion exchange resin (H+) and Sept,ember 1963 A. Falaschi, J. Adler, and H. G. Khorana 3081 the total effluent and washings evaporated after addition of pyridine (1 ml). To the residue was added triethylamine (0.05 ml) and pyridine (2 ml) and the solution was re-evaporated with vacuum from an oil pump. The residue was rendered anhydrous by repetition of evaporation after addition of dry pyridine. Finally, dry pyridine (0.5 ml) and acetic anhydride (0.2 ml) were added and the sealed reaction mixture kept in the dark at room temperature for 4 hours. Water (2 ml) was then added and the total solution kept for 2 hours at room temperature. It was then evaporated under reduced pressure to an oil which was extracted with dry ethyl ether several times. The insoluble polynucleotide material was obtained as a fine solid deposit on the wall of the round bottom flask. It was dissolved in water and the aqueous solution was lyophilized. The solid residue was made up to 0.2 ml with water. Resistance of Polynucleotides Bearing Terminal S'-0-Acetyl Grou.ps to E. coli Phosphodiesterase-An exonuclease purified from E. coli has been shown by Lehman (18) to degrade deoxy- ribopolynucleotides in a stepwise manner from the end bearing a 3'.hydroxyl group. The reaction produces deoxyribonucleo- side 5'-phosphates until the chain length is reduced to the dinucleotide (18). In a control experiment, 0.17 pmole of the tetranucleotide d-pTpTpTpT was incubated at 37" in a 0.2-ml incubation mixture in the presence of 0.02 ml of 1 M Tris buffer (pH 7.5), 0.02 ml of 0.1 M magnesium chloride, and 100 units (18) of enzyme. Degradation to d-pT and d-pTpT was com- plete in under 2 hours as determined by paper chromatography of aliquots in descending ethyl alcohol-O.5 M ammonium acetate buffer, pH 3.8 (7:3, v/v). Incubation of the 3'-O-acetyl d- pTpTpTpTpT under identical conditions up to 4 hours showed complete resistance of the oligonucleotide to the enzyme. In a second experiment the use of a S-fold higher concentration of the enzyme preparation under the above conditions showed like- wise the absence of a,ny degradation. On the other hand, the degradation of oligonucleotides bearing 3'-hydroxyl groups pro- ceeded normally in the presence of the 3'-0-acetyl derivatives, showing that the latter were not inhibitory. dssay of RNA Polymerase-The reaction mixture was exactly the same as that described by Chamberlin and Berg (lo), except that synthetic deoxypolynucleotides usually replaced DNA. When DNA was used as primer, the C'4-RNA synthesized was measured after precipitation with perchloric acid, exactly as described by Chamberlin and Berg (10). When synthetic deoxypolynucleotides were used, a new assay was devised in order to be able to detect any low molecular weight ribopolynucleotides that might be too small to precipitate in acid. The reaction was stopped with 0.02 ml of concentrated ammonium hydroxide and the entire mixture was then deposited in one pipetting onto the origin of a 2.5 x 57-cm strip of DEAE- cellulose2 paper (Whatman DE-20). Descending chromatog- raphy was carried out in 0.3 M ammonium formate for 2$ hours. Under these conditions polynucleotides as small as ribotetra- adenylate remain at the origin, whereas unused nucleoside tri- phosphates move away. The strip was then left to dry in air. The area containing the C'4-ribopolynucleotides (from 2.5 cm below to 2.5 cm above the origin) was cut out, folded in half at the origin, placed folded up in a scintillation vial containing sol- vent (3 g of 2,5-diphenyloxazole and 100 mg of 1,4-bis-2-(5- phenyloxazolyl)benzene per liter of reagent grade toluene), and 2 The form DEAE- refers to diethylaminoethyl-. counted at 2-80 (875 volts) and window settings of 10, 50, and 100 in a Packard Tri-Carb liquid scintillation spectrometer. Determination of Size of Ribopolynucleotides-After the area containing the C%ibopolynucleotide had been counted as de- scribed above, it was cut into small pieces and incubated in 2 ml of 0.3 N NaOH for 20 hours at 37" to hydrolyze the ribopoly- nucleotide. The liquid was then filtered and chilled in ice. It was neutralized by adding slowly small amounts of dry Dowex 50 (H+) resin; after each addition, about 5 minutes with occasional stirring were allowed before the pH was measured. When the pH reached 7, 5 ~1 of 1 N i\`aOH were added in order to avoid any dephosphorylation which could occur if the pH became less than 7 during the subsequent manipulations. The supernatant liquid, combined with 0.1 N ammonium hydroxide washings of the resin, was concentrated by lyophilization and spotted on a strip, 2.5 cm X 57 cm, of DE%E-cellulose paper. Descending chromatography in 0.2 .M ammonium formate was carried out for 6 hours until the front reached the end of the paper. This served to separate added (as markers) adenosine, adenosine 3'-phosphate, and adenosine 2'(3') ,5'-diphosphate. The dried strip was cut at l-cm intervals and these pieces were put into scintillation vials and counted as described above. The ratio of radioactivity in adenylic acid to adenosine or in adenylic acid to the adenosine diphosphate was considered the average size of the ribopolynucleotide. RESULTS De~ypolytkyn~idylate as Template fw Synthesis of Ribopolyadenylate E$ect of Chain Length on Rate of SynthesisThe initial rate of incorporation of adenylate varied with the size of the deoxypoly- thymidylate at saturating concentrations of each polymer, as shown in Fig. 1. With dTB no activity was observed under the conditions used, and with dT4 there was occasional activity. But dT5 always brought about significant, although small, incorporation of adenylate. With further increase in size the effectiveness of the polymers then rose, at first slowly up to dT7, then with a big leap upward between lengths of 8 and 11. The activity reached i FIG. 1. Effect of chain length on the activity of deoxypoly- thymidylate. The experimental procedure for measuring the incorporation of adenylate from C"-ATP is described under "Assay of RNA Polymerase." 3082 Synthetic Deoxypdynucleotides as Templates for RNA Polymeruse Vol. 238, No. 9 ThBLE I Ejiciency of DNA an.d deoxypolythymidylate in stimulating RNA polymerase The experimental procedure is described under "Assay of RXJA Polymersse." Incorporation of nucleotide Polymer CTP, GTP, UTP, and Only Cl"-ATP W-ATP present present Calf thymus DI'iA. Heated calf thymus DNA*. Deoxypolythymidylate, 1 M fraction. dT,,. . mjLmoles/I0 min 14.1 0.82 4.8 3.05 20.0 34.0 * Thymus DNA was heated by immersion in a boiling water bath for 10 minutes, then chilled in ice. a maximum at dTld and d'l'lj. Table I shows that dT1( is actu- ally more active for ribopolyadenylate synthesis than DNA is for the synthesis of RNA in the presence of all four nucleoside triphosphates. Hurwitz et al. (7, 11) had already &own that the 1 IVI triethyl- ammonium bicarbonate fraction referred to above is active for the synthesis of ribopolyadenylate. We have now confirmed the activity of the same 1 ~1 fraction and have found it to be about 60% as a&ive as dT,, (Table I). This lesser activity is probably due to the presence of less active polymers or inhibitors in the mixture. As mentioned above, the 1 M fraction contains polymers of various sizes larger than dTn and also oligonucleo- tides linked together by pyrophosphate bridges. It seems very likely that such pyrophosphate compounds are inhibitors for the priming action of polynucleotides. RKA polymerase will catalyze the synthesis of ribopoly- ndenylate when DNA is present together with ATP as the only nucleoside triphosphate (10). Denat,ured DNA is a preferred primer for this activity (24).3 Table I shows a confirmation of these results and a comparison of DNA with the activity of the 1 M fraction and dT14. For the synthesis of ribopolyadenylate tlT14 is the most active polymer. In order to be sure that the comparison of the results with polynucleotides of different sizes was meaningful, it was neces- sary to check that the primer did not undergo degradation during the incubation Fith IZKLY polymerase (see also below). W-Labeled dT7 was used in the usual reaction mixture except that ATP was omitted. The mixture was incubated as usual and then put on a DEAE-cellulose carbonate column (0.6 x 27 cm) and cluted with a linear gradient of triethylammonium bicarbonate from 0 to 0.48 >\I (total volume, 500 ml). More than 99.7 7` of the radioactivity appeared as a single peak in the posi- tion corresponding to dTs. This ruled out the possibility of detectable breakdown of the polymer by the enzyme preparation. Effect of Chain Length on Saturoling Concentration of Tern- plate--l saturating concentration was determined for each poly- mer that stimulated the synthesis of ribopolyadenylate in the experiment of Fig. 1. For dTi, dTll, and dT14 the results are plotted in Fig. 2 according to the method that Lineweaver and Burli have used for substrates (25). The polymer concentra- tions which gave half-masimal rabes with dT7, dTu, and dT14 were found to be 50 X 10e6, 20 X lOW, and 2.0 X 10-e RI, re- spectively. Increasing the size of the polymer not only incremes the maximal velocity of the reaction (as shown also by Fig. 1) but decreases very strikingly the concentration of polymer re- quired for half-saturation. Apparently the affinity of the deoxy- polythymidylate for the enzyme increases markedly with size be- tween dTr and dT14. Size of Ribopolyndenylate Ii'ormed-The C%ibopolyadenylate formed in the presence of deoxypolythymidylate of various sizes was hydrolyzed with sodium hydroxide to CWadenosine, U4-adenosine 2'(3')-phosphate, and a W-material that resem- bled adenosine 3',5'-diphosphate but was not further char- acterized. The details are described under "Experimental Pro- cedure." The ratio of radioactivity in adenylic acid to adenosine and in adenylic acid to the adenosine diphosphate was taken to be the average size of the ribopolyadenylate. Table II lists the results for the size of the products formed from dTi, dT*, dT9, dTll, and dT14; for smaller thymidine oligonucleotides too little product was available to provide significant results. The estimate of the chain length of the products is only approximate, owing to t'he inaccuracy of counting on DEAE paper the small amounts of radioact,ivity in the adenosine and in t,he adenosine diphosphate region. The most striking feature of the results is that the product is of a much larger size than the deoxypoly- thymidylate. No marked differences are apparent between the sizes of the products obtained with different sized deoxypoly- thymidylates. Evidence for Noninvolvement of Terminal S'-Hydroxyl Groups of Polythymidylate in Ribopolyadenylate Formation-During chromatography of the total alkaline hydrolysate of the ribopoly- adenylate (see above), no significant amount of radioactivity FIG. 2. Reciprocal plot of the effect of concentration of deoxy- polythymidylate on the incorporation of adenylate. il compares dT7 and dT11; B compares dTI, and dTl4. The concentration of polymer is expressed as millimicromoles of polymer (not nuc- leotide equivalents) per ml; ZJ, millimicromoles of adenylate incorporated from ATP per 10 minutes, as measured by the DEAF,-cellulose paper assay described under "Assay of RNA - _ . Yolymerase." 3 M. Chumberlin and P. Berg, unpublished observations. September 1963 A. Falaschi, J. Adler, and H. G. Khoran.a 3083 remained at the origin of the chromatogram where the deoxyribo- polythymidylate remains adsorbed. This result indicated that ribopolyadenylate synthesis was not initiated by esterification of the terminal 3'-hydroxyl group in deoxypolythymidylate to form a phosphodiester linkage. Such a linkage would have been resistant to alkaline hydrolysis and a product of the general structure (Diagram I) would have resulted. T T T A `P DIAQRAM I To test further whether the 3'-hydroxyl group is essential in the synthesis of ribopolynucleotides, 3'-O-acetyl-dT11 was pre- pared for testing as a primer. Table III (Lines 1 and 3) shows that 3'-0-acetyl-dT11 is about as active as dT,,. The concen- tration of 3'-0-acetyl-dTn required to half-saturate the enzyme was very similar to the concentration already found for dT1r. This diminished the possibility that a small amount of unacety- lated dTll was the active component in the 3'-0-acetyl-dTu prep- aration. Furthermore, 3'-0-acetylthymidylate was found to be T.~BLE II Size of ribopolyadenylate formed from deoxypolylhymidylate templates The experimental procedure is described under "Size of Ribo- polyadenylate Formed." I Radioactivity I Chain length c.p.m. dT7 29 2,550 96 dTe 80 10,100 135 dT, 70 4,680 32 dT,l 34 3,590 86 dT1, 114 9,080 / 63 73 27 50 12.5 74 100 67 146 107 101 46 74 79 144 111 TABLE III Acetylaled deozypolythymidylate as template fo~ RXA polymerase Treatment with E. coli phosphodiesterase was carried out at 37" in a 0.15-ml react,ion mixture containing 100 units (18) of enzyme, 0.02 ml of 1 M Tris buffer (pH 7.5)) 0.01 ml of 0.1 M mag- nesium chloride, and 0.4 pmole of nucleotide equivalent of poly- nucleotide. Then 0.02.ml aliquots of this were added to an RNA polymerase reaction mixture without prior inactivation of the diesterase, since it could be shown that the presence of diesterase does not interfere with the synthesis of ribopolyadenylate. Polymer Treatment with ' Incorporation of phosphodiesterase / adenylate from ATP I mpnmlesj30 win dT1 ,... _. _. - 2.30 dTII.. + I 0.04 3'-0.acetyl-dTI1.. - 2.37 3'-0-:tcetyl-dT11, + i 1.32 TABLE Iv Deoxypolycytidylate OS template for synthesis of ribopolyguanylate The experimental procedure is described under "Assay of RNA Polymerase." Polymer Incorporation of guanylate from GTP 1 Incorporation of j adenylate from ATP dCs dCs dCs dCs dClo dTs rrqmole: 10 viin <0.05 (0.05 <0.05 0.15 0.17 3.2 stable when incubated with the RS*4 polymerase preparation; this served to exclude any acylase activity, In order to ensure that none of the unacetylated dTn was present, the preparation of t,he acetylated polynucleotide was preincubated with the E. co/i phosphodiesterase. Acetylation of the 3'-hydroxyl group in thymidine oligonuclcotides confers resistance toward this enzyme (see above). Table III (Lines 2 and 4) compares dTi, and 3'-0-acetyl-dTn after treat#ment with phosphodiesterase. This enzyme abolishes nearly all of the activity of dTll whereas most of the activity of 3'-O-acetyl-dT11 remains resistant to phosphodiesterase. It may be concluded that 3'-0-acetyl-dT,, is active for the synthesis of ribopolyadenylat'e, and that a free 3'-hydroxyl group of deoxypolythymidylate is not essential for ribopolyadenylatc synthesis to occur. It follows that addition of adenylate to the 3'-hydroxyl end of deoxypolythymidylate is not a necessary part of ribopolyadenylate synthesis. Deoxypolycytidylate as Template for Synthesis of Rihopolyguarqlate Incubation of deoxypolycytidylate with RNA polymerase and GTP led to the formation of ribopolyguanylate, and the rate of the reaction depended on the size of the deoxypolycytidylate (Table IV). Whereas a chain length of 5, 6, or 8 proved too small to be effective, dC9 and dCla showed significant activity. The dependence of the reaction on the concentration of dClo is shown in Fig. 3; half-saturation occurred at 5.4 x 10-C Y. A comparison of the deoxypolycptidylate series with the deoxypoly- thymidylate series shows that lower homologues are more effec- tive in the latter series (Table IVj but a lower concentration of the deoxypolycytidylat,e is required for saturating the enzyme. Deoxypolycytidylate of much higher molecular weight than used here, prepared from the deoxypolycytidylate-deoxypolyguanylate polymer (26), has been shorrn by Chamberlin and Berg to be active for the synthesis of ribopolyguanylate (27). Experinlents with Deoxypolyguanylak and Deoxypolyadenylate In the deoxypolyguanylate series, only dGs was tested. It proved to be inactive for the incorporation of nucleotide from CTP. This may be explained by the finding that even small homologues of deoxypolyguanylute have a high tendency to form aggregates of very large molecular weight (28). Chamber- lin and Berg (27) have reported that deoxypolyguanylate pre- pared from the deoxypolycytidylate-deoxypolyguanylate poly- mer is also inactive for the incorporation of cytidylate. 3084 Synthetic Deoxypolynucleoticles as Templates for RNA Polynaerase Vol. 238, No. 9 Fro. 3. Reciprocal plot of the etl'ect of concentration of dClo on the incorporation of guanylate. The experimental procedure is described under "Assay of RSA Polymerase." The concentra- tion of polymer is expressed as millimicromoles of polymer (not nucleotide equivalents) per ml; u, millimicromoles of guanylate incorporated from GTP in 10 minutes, as measured by the DEAE- cellulose paper assay described under "Assay of RNA Polymer- me." TABLE 1 Specificity of incorporation of nucleotide The experimental procedure is described under "Assay of RNA Polymerase." Polymer Nucleotide incorporated ATP only / GTP only I I CTP only UTP only Poly dT, 1 M fraction dClo Poly dA, 1 M fraction dGs ?npmoles/l0 min 8.7 i