PHOSPHOROLYSIS OF OLIGORXBONUCLEOTIDES BY POLYNUCLEOTIDE PHOSPHORYLASE BY MAXINE F. SINGER* (From the National Institute of Arthritis and Metabolic Diseases, National Institutes of Health, United States Public Health Service , Bethesda, Maryland) (Received for publication, December 23 , 1957) The enzyme polynucleotide phosphorylase catalyzes the over-all reaction described in Equation 1. n nucleoside-PP . Mg* (nucleoside-P), + n inorganic P (1) This reversible reaction has been demonstrated With enzyme preparations from Azotobacter trinelundii (1, 2) and Escherichia coli (3). In the forward direction the enzyme catalyzes the formation of polyribonucleotides sim- ilar in structural details to natural RNA1 (5-8). In the reverse reaction polyribonucleotides are phosphorolyzed to yield nucleoside diphosphates. The phosphorolysis of synthetic polymers made by the enzyme and the phosphorolysis of ribonucleic acids isolated from various natural sources have been studied by Ochoa and coworkers with the A. VineZundii enzyme (2, 9), by Littauer and Kornberg with a preparation from E. coli (3), and by Heppel.2 Highly polymerized RNA preparations were phosphorolyzed at slower rates than the synthetic polymers; however, commercial RNA and RNA "core," the limit polynucleotides obtained after exhaustive * Research Fellow of the National Institute of Arthritis and Metabolic Diseases, United States Public Health Service. 1 The following abbreviations are used: adenosine and uridine units are repre- sented by A and U, respectively; 5`-diphosphates of adenosine and uridine, ADP and UDP ; 5'-nucleoside diphosphate, NDP; 5'-monophosphates of adenosine and uridine, AMP and UMP; 5`-triphosphate of adenosine, ATP; ribonucleic acid, RNA; polyadenylic acid, poly A; polyuridylic acid, poly U; mixed polymer of adenylic and uridylic acids, poly AU; tris(hydroxymethyl)aminomethane, Tris; reduced diphos- phopyridine nucleotide, DPNH; inorganic orthophosphate, Pi. Small polynucleo- tides are designated by a system proposed by Markham and Smith (4). A phosphate group is designated by "p"; when placed to the right of a nucleoside symbol, the phosphate is esterified at C-3' of the ribose moiety; when placed to the left of the nucleoside symbol, the phosphate is esterified at (2-5` of the ribose moiety. Thus, pApA is a dinucleotide with 1 phosphate monoesterified at C-5` of an adenosine residue and a phospho diester bond between C-3' of that same adenosine residue and (2-5' of the other adenosine group. The symbol -cyclic-p is used to designate a terminal 2',3'-cyclic phospho diester moiety on a polynucleotide. Thus, UpU- cyclic-p is a dinucleotide with a 2',3'-phosphate at the terminal ribose. 2 Cited by Littauer and Kornberg (3) and Ochoa (9). 211 212 PHOSPHOROLYSIS OF OLIGONUCLEOTIDES digestion of RNA with pancreatic ribonuclease, were attacked very slowly, if at all. In the light of these data it was not possible to define any require- ments, by polynucleotide phosphorylase, for specific structural details in the polynucleotide substrates. The preparations used were not of uni- form, known chain length and, in addition, the end group structures of the natural RNA preparations have not been clearly established. Heppel and coworkers recently described methods for the preparation of several homologous series of well characterized oligoribonucleotides (7, 8, 10). This made it possible to investigate the specificity of the phosphoro- lysis reaction. The experiments to be described in this paper indicate that polynucleotides with C-5' phospho monoester end groups are readily phos- phorolyzed although those bearing (3-3' phospho monoester end groups are resistant to enzymic attack. The phospho monoester at (2-5' is not es- sential for activity, however, since oligonucleotides with no monoesterified phosphate groups, such as trinucleoside diphosphates, are phosphorolyaed. In addition, it was found that dinucleotides and dinucleoside monophos- phates are not attacked by polynucleotide phosphorylase, and these com- pounds accumulate as resistant end products when the phosphorolysis of the larger oligonucleotides is studied. A preliminary report of these findings has been made (11). EXPERIMENTAL Materials-Polynucleotide phosphorylase was purified from E. coli by Dr. R. J. Hilmoe according to the procedure of Littauer and Kornberg (3). The fraction used is described as Ethanol I (3) and was dialyzed before use. The preparation contained 1 mg. of protein per ml. and had a specific ac- tivity of 15, determined by the "exchange" assay (Assay C) (3). The preparation of A. vinelandii polynucleotide phosphorylase was kindly sup- plied by Dr. s. Ochoa. The fraction was an eluate from Ca3(P04)z gel (12), contained 7.4 mg. of protein per ml., and had a specific activity of 40 as measured with the "exchange" assay (Assay 2) (2). Crystalline bovine pancreatic ribonuclease was a commercial preparation (Armour, lot No. 1044). Phosphomonoesterase was fractionated from human seminal plasma (4) after a preliminary treatment with protamine to remove nucleic acid material. The specific activity of this preparation was 3.7 X lo3 units per mg. of protein (1 unit of enzyme liberates 1 pmole of inorganic phosphate per hour per 1.2 ml. of a reaction mixture containing 26 pmoles of AMP, in acetate buffer, pH 5.2). This enzyme preparation dephosphoryl- ated dinucleotides more slowly than mononucleotides. Therefore, for the preparation of dinucleoside monophosphates a great excess of enzyme was used. Even more enzyme was required to dephosphorylate oligonucleo- tides of chain length greater than 2. The preparation used in these experi- M. F. SINGER 213 ments was relatively free from phosphodiesterase activity, but, when it was used in the high concentrations required to dephosphorylate oligonu- cleotides, appreciable diesterase activity was noted. The nuclease from guinea pig liver nuclei, which liberates small polynucleotides with C-5' phosphate end groups from poly A, has been described by Heppel and co- workers (10) .a A commercial preparation of lactic dehydrogenase (Worth- ington, crystalline) which was contaminated with pyruvate kinase was used as a source of both of these enzymes. Myokinase (63 units per ml., 0.8 mg. of protein per ml.) was supplied by Dr. B. L. Horecker and was pre- pared according to the procedure of Colowick (13). Polymers were prepared from nucleoside diphosphates with the poly- nucleotide phosphorylase of E. coli or A. winelandii according to the pro- cedure of Grunberg-Manago and coworkers (2), but the method of isolation was slightly modified. After polymer formation at pH 8, the reaction mix- tures were adjusted to pH 7 with acetic acid. Polymer was precipitated with 3 volumes of ethanol, separated by centrifugation, dissolved in water, and reprecipitated in the same manner. The twice precipitated polymer was dissolved in water and the solution was deproteinized by shaking with 0.25 volume of chloroform and 0.1 volume of isoamyl alcohol (14). The aqueous solution of polymer was then dialyzed against cold, running, dis- tilled water for 3 days and lyophilized. A sample of poly A prepared with the polynucleotide phosphorylase of Micrococcus Zysodeikticus (15) was generously donated by Dr. R. F. Beers. All the mononucleotides used were commercial preparations (Sigma). Phosphoenolpyruvic acid was prepared by Mr. William E. Pricer, Jr., and DPNH was donated by Dr. B. L. Horecker. Paper Chromatography and Paper Electrophog"esis-Severa1 systems were used for the separation of mononucleotides and oligonucleotides on paper. Descending chromatography was carried out with the following solvent systems: System 1, isopropanol-water (70: 30, v/v) with NHI in the vapor phase (16) ; System 2, isobutyric acid-1 M NHIOH-0.2 M ethylenediamine- tetraacetate (100: 60:0.8, v/v/v) (17). For Solvent 1, Whatman No. 3MM paper was used and for Solvent 2, Whatman No. 3MM or Whatman No. 1. Electrophoretic separations (referred to as System 3) were carried out ac- cording to Markham and Smith (16) on strips (57 X 10 cm.) of Whatman No. 3MM paper saturated with 0.05 M ammonium formate buffer, pH 3.5. A potential of lo00 volts was applied across the paper. Purine- and pyrimidine-containing compounds were located on the paper strips with an ultraviolet light which was also used to photograph the strips. When these methods of separation were used in a preparative way or to obtain materials for enzymic treatment, the purine and pyrimidine derivatives *We thank Dr. R. J. Hilmoe and Dr. L. A. Heppel for this preparation. 214 PHOSPHOROLYSIS OF OLIGONUCLEOTIDES were eluted with water. When quantitative elution of the material was desired, the ultraviolet-containing areas were eluted with 0.01 N HCl for 6 hours at room temperature, and the concentration of the compound was determined by measuring the absorption of the eluate at an appropriate wave length. From the adjacent region of the paper strip, an area of identi- cal size was cut out and eluted in order to correct for the ultraviolet-ab- sorbing material present in the paper itself. Assay Procedures-In the procedure used (3), the phosphorolysis of polynucleotides was carried out in the presence of inorganic P32. The resultant labeled nucleoside diphosphate (see Equation 1) was separated from inorganic Pa2 by adsorption onto charcoal and its radioactivity was measured. The reaction mixtures (0.125 ml.) contained the oligonucleo- tide or polymer, enzyme, 5 pmoles of Tris buffer, pH 8.0, 0.5 pmole of MgC12, and 3.2 pmoles of P32-labeled sodium potassium phosphate buffer, pH 7.4. After incubation at 37", the reaction was stopped by adding 0.1 ml. aliquots to 1.0 ml. of cold, 2.5 per cent perchloric acid. Acid-washed Norit A (0.1 ml. of a 10 per cent suspension, w/w) was added to adsorb the nucleotides. After 10 minutes in the cold, the suspension was centrifuged and the charcoal was washed three times with 2.5 ml. portions of water. The charcoal was then suspended in 0.8 ml. of 50 per cent ethanol contain- ing 0.3 ml. of concentrated NH,OH per 100 ml. An aliquot of this sus- pension (usually 0.1 ml.) was placed on a copper planchet, dried, and the radioactivity determined with a thin window, gas flow counter. The total counts per minute incorporated into charcoal-adsorbable material were de- termined and from the specific radioactivity of the inorganic Pa2 the micro- moles of phosphate incorporated were calculated. A self-absorption factor of 1.15 (3) was applied. Two control incubations, one containing no sub- strate and one containing no enzyme, were generally carried through the whole procedure with each experiment, and the results presented have been corrected for the small amount of radioactivity adsorbed onto the charcoal from these samples. When the products of phosphorolysis were to be isolated, the reactions were carried out in the same manner but on a larger scale. Nucleotides were eluted from charcoal by four treatments with 0.8 ml. of ethanolic ammonia; the ethanol supernatant fluids were pooled and concentrated before chromatography. The phosphorolysis of polynucleotides was also measured by determining the nucleoside diphosphate formed with the spectrophotometric procedure of Kornberg and Pricer (18). The application of this method to the de- termination of polynucleotide phosphorylase activity has been described by Ochoa and Heppel (6). Myokinase was added to the assay system to reconvert to ADP any AMP and ATP that had been formed in the original M. F. SINGER 215 incubation as a result of myokinase activity in the polynucleotide phos- phorylase preparation. UDP was also measured by this method (19). All optical measurements were carried out with a Beckman model DU spectrophotometer . Preparation of Oligonucleotides-Five series of homologous oligonucleo- tides were prepared in order to study the specificity of polynucleotide phos- phorylase. Each series included compounds of 2, 3, and 4 nucleoside residues and the structure of a representative member of each group is illustrated in Fig. 1. In every compound studied, the internucleoside links were 3f, 5f-phospho diester bonds. The oligonucleotides used in this work were recently investigated by Heppel and coworkers (7,8, lo), who isolated them from biosynthetic polymers and characterized them by methods developed by Markham and Smith (4), Volkin and Cohn (20), and others. The compounds in Group I contained adenosine as the only nucleoside moiety and each of the three oligonucleotides had a monoesterified phos- phate at C-5' of the terminal nucleoside (7, lo). The structure of the trinucleotide, pApApA, is shown in Fig. 1. These compounds were de- scribed and characterized by Heppel and coworkers (7, 10) and are formed by the action of a nuclease from guinea pig liver nuclei on synthetic poly A. A typical preparation was carried out in the following way: The incuba- tion mixture (30 ml.) contained 400 pmoles of MgC12, 800 pmoles of phos- phate buffer, pH 7.2, 80 mg. of poly A (M. Zysodeikticus), and 5 ml. of enzyme containing about 30 mg. of protein. After 6 hours of incubation at 37", toluene was added to inhibit bacterial growth. After 24 hours of incubation the mixture was cooled, deproteinized with chloroform (14), and concentrated by lyophilization. The resulting solution was divided into eight equal portions and each was applied as a thin band to filter paper and chromatographed for 64 hours in System 1. The di-, tri-, and tetra- nucleotides (pApA, pApApA, and pApApApA) were eluted from the papers and the eluates were concentrated at 40" in a stream of air. The concentra- tion of oligonucleotide was estimated by measuring the ultraviolet absorp- tion of the eluate at 257 mp and applying the extinction coefficient for adenylic acid.4 The dinucleotide (pApA) that was obtained in this way was contaminated with ADP and inorganic P and was further purified by paper electrophoresis. It was assumed for this work that the molar extinction coefficient of an oligo- nucleotide is approximated by the sum of the molar extinction coefficients of its constituent nucleotides. For example, the concentration C of ApApUp in micro- moles of oligonucleotide per ml., when the absorption is measured at pH 2, is given by C = Ab~.~~~/(2(15.1) + 10.0), where the molar extinction coefficients for AMP and UMP are 15,100 and 10,000, respectively. Recent unpublished experiments by the present author indicate that, after alkaline hydrolysis, the absorption of pApApA at 257 mp and pH 7 increases by approximately 15 per cent. Enzymic hydroIysis of dinucleotides in the deoxyribose series also results in an increased absorption (21). 216 PHOSPHOROLYSIS OF OLIGONUCLEOTZDES The compounds of Group I1 were derived from those of Group I by the removal of the monoesterified phosphate at (3-5' by human seminal plasma phosphomonoesterase (7, 10). The structure of the trinucleoside diphos- T*e C I C M2f"l"" F@w Trinucleotide of Qvup I 4w Trinucleoside diphosphate of Oroup 11 P P. UpVpO-cyc=C-P Trinucleatids of mnlp v FIG. 1. Outline structures of the trinucIeotides and trinucIeoside diphosphates phate (ApApA), which was derived from pApApA, is shown in Fig. 1. The di-, tri-, or tetranucleotide of Group I was incubated with monoesterase under the conditions described by Heppel and coworkers (7), and the cor- M. F. SINGER 217 responding nucleoside derivatives were separated by chromatography in System 1 (10). The products were obtained from the paper strips in the manner described for the compounds of Group I. The three oligonucleotides in Group I11 were obtained by exhaustive digestion of poly AU with ribonuclease (8). The structure of the trinucleo- tide, ApApUp, is shown in Fig. 1. These oligonucleotides contained vary- ing numbers of adenosine residues but each had a terminal uridine unit with a monoesterified phosphate at its C-3'. Thus the homologous di- and tetranucleotides were ApUp and ApApApUp, respectively. The electrophoretic and chromatographic properties of these compounds and the methods of identification have been recorded by Markham and Smith (4) and Volkin and Cohn (20). In a typical preparation, 15 mg. of poly AU were incubated with 0.6 mg. of ribonuclease for 16 hours at 37" in 0.05 M Tris buffer, pH 8 (total volume equal to 1.7 ml.). The mixture was deproteinized as described above (14), and the aqueous layer was concen- trated, applied to paper, and chromatographed in System 1. The bands obtained were eluted and the absorption of the eluates at 260 mp was de- termined. The sum of the extinction coefficients for each mononucleotide residue was used to estimate concentration of the oligonucleotide? The dinucleoside monophosphate ApU, trinucleoside diphosphate ApApU, and tetranucleoside triphosphate ApApApU of Group IV were obtained from the corresponding oligonucleotides of Group 111 by treat- ment with seminal phosphomonoesterase, exactly as described for the preparation of Group 11 from Group I. The properties of the compounds in Group IV are described in the literature (8, 4, 20) and differ from those of Group I1 in that the terminal nucleoside is uridine (see Fig. 1). The concentrations of these compounds were estimated as described for Group 111. The last homologous series of oligonucleotides (Group V) was obtained from the controlled digestion of poly U with small amounts of ribonuclease (7). Each member of the group contained exclusively uridine residues with a cyclic-terminal phospho diester moiety. The structure of the cyclic terminal trinucleotide, UpUpU-cyclic-p, is shown in Fig. 1. These com- pounds were separated by chromatography in Solvent 1 and were eluted from the papers with water. By applying the extinction coefficient for uridylic acid,' the concentrations of the oligonucleotides were determined by measuring the absorption of the eluates at 262 mp. Results The trinucleotide and tetranucleotide of Group I, which had a phospho monoester group at C-5', were readily phosphorolyzed by the polynucleo- tide phosphorylases of E. coli and A. Vinektndii (Table I). The compounds 218 PHOSPHOROLYSIS OF OLIGONUCLEOTIDES pole per ml 0.3 0.3 0.4 0.4 0.4 0.3 0.9 of Group 111, which contained a phospho monoester moiety at C-3', were, however, resistant to phosphorolysis by either of the enzymes (Table 11). The data in Table I1 suggested that ApApApUp was very slowly cleaved by the A. vinehdii enzyme, but this was not confirmed by other studies with both enzyme fractions. Incubation ww carried out as in Table II but for a period of 24 hours; chromatographic investigation of the reaction TABLE I Phosphorolysis of Compounds in Group I and II I I E. coli experiment Substrate 1 Substrate I 1 Rateof concen- radi- phospho- tration. actlvlty+ rolysist Group No. ~ I I1 1 Poly A c.9.m. 0 3470 8910 0 1899 6080 4620 0.0 0.6 1.4 0.0 0.3 1 .o 9.3 A. &clundii experiment Substrate concen- tration* mole pa ml. 0.6 0.6 0.7 0.7 0.7 0.9 0.9 'otal radi. mctivityt c.p.m. 130 3,700 12,780 18 2,280 6,380 3,880 Rate of hosphor- olysist 0.0 0.7 2.6 0.0 0.5 1.3 7.4 * Concentrations are expressed in terms of adenosine equivalents. To obtain the t Total number of counts per minute adsorbed onto charcoal. $ Micromoles of phosphate incorporated into charcoal-adsorbable compounds per hour per mg. of protein. In the experiment with the E. coli enzyme, each tube con- tained 786,000 c.p.m. as Pial and 26 y of protein except for the poly A incubation mixture, which contained 2 y of protein; incubation time, 1 hour. In the experiment with the A. vinelandii enzyme, each tube contained 556,000 c.p.m. as Pi32 and 14 y of protein except for the poly Aincubation mixture, which contained 1.5 y of protein; incubation time, 2 hours. The other components of the reaction mixtures and the procedures are described under the section on methods. concentration as oligonucleotide, divide by the chain length. mixtures failed to reveal any significant breakdown of ApApApUp. The remaining data in Tables I and I1 show that, although the C-3' phospho monoester moiety inhibits the phosphorolysis of an oligonucleotide, the C-5' phospho monoester group is not a structural requirement for enzymic action. Thus ApApA and ApApApA (Table I) as well as ApApU and ApApApU (Table 11) were phosphorolyxed at rates that do not differ greatly from those found with pApApA and pApApApA. It was also found that the three compounds in Group V, which contain the terminal cyclic-p, were completely resistant to enzymic attack. M. F. SINGER 219 Only qualitative conclusions may be drawn from the relative rates of phosphorolysis given in Tables I and 11. InsufEcient amounts of material made it impossible to conduct the experiments at saturating substrate concentrations and only the following limited conclusions concerning rates can be drawn : (1) with equivalent concentrations of active oligonucleotide, the tetranucleoside derivative is phosphorolyzed more rapidly than is the TABLE I1 Phosphorolysis of Compounds in Group 111 and IV Group No. I Ir IV Substrate Poly A E. coli experiment Substrate concen- tration* mots )er ml. 0.4 0.3 0.3 0.3 0.3 0.3 0.9 *ob! radi. lac tivi tyt c.9.m. 0 18 110 0 644 2600 1781 Rate of phospho- rolyaist 0.0 0.0 0.0 0.0 0.2 0.8 10.4 A. vinelandii experiment Substrate concen- tration. mole 9.w ml. 0.7 0.8 0.8 0.6 0.6 0.3 0.9 'otal radi lactivityt c.P.m. 399 340 1682 115 4640 3145 3365 * Concentrations are expressed in terms of nucleoside equivalents. 1 the concentration, as oligonucleotide, divide by the chain length. t Total number of counts per minute adsorbed onto charcoal. Rate of phospho- rolysist 0.0 0.0 0.1 0.0 0.4 0.3 5.6 obtain $ Micromoles of phosphate incorporated into charcoal-adsorbable compounds per hour per mg. of protein. In the experiment with the E. coli enzyme, each tube con- tained 544,000 c.p.m. as Pisa and 20 y of protein except for the poly A incubation mixture, which contained 1 y of protein; incubation time, 1 hour. In the experiment with the A. winelandii enzyme, each tube contained 628,000 c.p.m. as Pia2 and 30 y of protein except for the poly A incubation mixture, which contained 1.5 y of pro- tein; incubation time, 2 hours. The other components of the reaction mixtures and the procedures are described under the section on methods. trinucleoside derivative, (2) the rates of phosphorolysis of corresponding members of the several groups are of the same order of magnitude, as for example pApApA, ApApA, and ApApU. Data on the phosphorolysis of poly A are included as a standard for comparison of rates and were obtained at a concentration of polymer which afforded a maximal reaction rate. Preliminary experiments carried out with pApApA indicated that a con- centration of 2.3 pmoles per ml. as adenosine units (0.8 pmole per ml. as trinucleotide) was sufficient to saturate the E. coli enzyme. When both were tested at saturating concentrations, the rate of phosphorolysis of pApApA exceeded that found for poly A by a factor of approximately 220 PHOSPHOROLYSIS OF OLIGONUCLEOTIDES 2-fold. The phosphorolysis of oligonucleotide attained a maximal rate at The production of nucleoside diphosphates as one of the products of the phosphorolysis of oligonucleotides was demonstrated by the spectrophoto- metric assay as well as by the chromatographic evidence discussed below. M inorganic P and required Mg++ ion. TABLE I11 Production of Nucleoside Diphosphates from Oligonucleotides Experiment No. Substrate Substrate concentration* lynolw per ml. 1.8 1.6 1.4 1.3 0.6 0.5 0.6 0.5 Pia incorporated into charcoal adsorbable I,rmok 0.00 0.13 0.14 0.03 0.04 0.00 0.02 0.04 Nucleoside diphosphate produced pm Is 0.00 0.14 0.19 0.03 0.04 0.00 0.02 0.04 * Concentrations are expressed in terms of nucleoside units. To obtain the con- centration, as oligonucleotide, divide by the chain length, The incubation mixtures were as described in the section on methods, except that the scale was doubled (final volume, 0.25 ml.). The E. coli enzyme was added in the following amounts: 20 y in Experiments 1 and 2, and 30 y in Experiment 3, After incubation for 2 hours at 37O, 0.1 ml. was treated as described for the standard assay to determine the micromoles of Pi** incorporated into charcoal-adsorbable compounds. Another 0.1 ml. aliquot was diluted to 0.5 ml. (Experiment 1) or 0.4 ml. (Experiments 2 and 3) with HtO, heated for 2 minutes at 100", centrifuged to remove the protein, and 0.30 ml. aliquots of the supernatant fluid were used to determine ADP or UDP. The components of the assay system (1.00 ml. in a cuvette with a 1.0 cm. light path) were 5 pmoles of MgC12, 0.4 pmole of phosphoenolpyruvate, 0.1 pmole of DPNH, 0.005 ml. of myo- kinase, and 0.005 ml. of lactic dehydrogenase containing pyruvate kinase. The data in Table 111 show that the amount of inorganic Ps2 incorporated into charcoal-adsorbable compounds was equal to the nucleoside diphos- phate formed. It was found that dinucleotides and dinucleoside monophosphates did not undergo phosphorolysis (Tables I, 11, and 111). This was true regard- less of the nature of the end group on the compound containing 2 nucleoside units. A similar conclusion was reached when the phosphorolysis of pApApA and pApApApA was studied as a function of time and the reac- tions were allowed to go to completion. Fig. 2 shows that poly A was M. F. SINGER 221 I I 0 0 A polymer - - - - I e 4 PAPAPAPA - PAPAPA - - W e - I 1 phosphorolyzed approximately 100 per cent; that is, the inorganic PS2 in- corporated into charcoal-adsorbable nucleotides when the reaction stopped was equivalent to the micromoles of polymer (expressed as adenine res- idues) originally added. With the trinucleotide, however, inorganic Pa uptake ceased when 1 pmole of phosphate had been consumed per 3 pmoles of adenine residues, or, to express this in another way, when 1 phosphate molecule had been incorporated per molecule of trinucleotide (Fig. 2). This suggested that the reaction proceeded according to the following equation PAPAPA + Pi + PAPA + ADP (2) and came to a halt when the limit dinucleotide wa.9 formed. Similarly, with the tetranucleotide, the incorporation of inorganic Paz ceased when 2 pmoles of phosphate were taken up per micromole of tetranucleotide (Fig. Z), again indicating that the dinucleotide is an end product of phosphoroly- sis. It should be mentioned that the accumulation of such a limit di- nucleotide during the phosphorolysis of poly A would have been too small to detect in this experiment. A dinucleotide tentatively identified as cyclic dianhydrodiadenylic acid was also tested as a substrate for E. coli polynucleotide phosphorylase. 222 PHOSPHOROLYSIS OF OLIGONUCLEOTIDES Neither the chemically prepared material (22) nor that obtained from mam- malian tissues (23) was phosphorolyzed by the Identification of Reaction Products-Chromatographic investigation of the reaction products confirmed the identification of the nucleoside diphos- phates and also showed that a dinucleotide or dinucleoside monophosphate, depending on the starting substance, accumulates during the phosphoro- lysis of trinucleoside or tetranucleoside derivatives. Reaction mixtures were treated as described under the section on methods and nucleotide material was eluted from the washed charcoal by successive treatments with ethanolic ammonia. The eluates were chromatographed in appropri- ate solvents, along with suitable markers, and the ultraviolet-absorbing spots were then eluted and treated in various ways to confirm their identity. In one typical experiment, the substrate was pApApApA and the charcoal eluate was chromatographed in solvent System 2. Fig. 3, a is a photograph of that chromatogram. The markers are on the left and the reaction mix- ture on the right. The most rapidly moving area (c) contained the di- nucleotide pApA and will be discussed in detail. The next area (b) cor- responds in Rp to ADP and the slowest moving area (a) to ATP. ATP arises from ADP owing to myokinase present in the E. coli enzyme prep- aration, and the AMP that is also formed is mixed with pApA in the fastest moving area. The ADP area (b) was eluted and its specific radioactivity determined. The data in Table IV give the specific activity of the ADP as well as that of the ADP and UDP isolated by similar techniques from the phosphorolysis of the other oligonucleotides. In each case, the spe- cific radioactivity of the nucleoside diphosphate was equal to that of the inorganic Paz present in the incubation mixture. Included in Table IV are the Rp values of authentic samples of the nucleoside diphosphates and those found for the isolated materials. As indicated in Table IV, this procedure did not afford quantitative recovery of the products. In the experiment with pApApApA (Fig. 3, a) the material with the highest Rp in Solvent 2 (Area c) was eluted and portions of it were subjected to analysis by Systems 1 and 3. A photograph which demonstrates the paper electrophoresis of this material is shown in Fig. 3, b. The known markers are indicated; the eluate from Area c of Fig. 3, a had the mobility of pApA. In System 1 the major portion of the material also behaved as pApA (Table V). Similar techniques were used to identify pApA as a product of the phosphorolysis of pApApA, and ApA as a product of the phosphorolysis of ApApA, ApApApA, ApApU, and ApApApU (Table V). The solvent system used for the initial separation of the dinucleotide or 6 We are indebted to Dr. Markham for a sample of the synthetic material and to Dr. Rall for a sample of the isolated compound. Jd. F. SINGER 223 dinucleoside monophosphate from the nucleoside diphosphate is shown in Table V; the material was eluted from the paper and then subjected to the other treatments indicaked. The dinucleotides and dinucleoside mono- PIG, 3. (a), uitraviolet photograph of paper chromatogram showing the products of phosphoroiysis of pApAphpA. The reaction mixture (0.15 mi. voIame) contained 5pinoles of Tris buffer, pfi 8,O.Spmole of MgC12,3.2 &moles of 1'132 containing 596,oOO c.p,ni., 20 y of E'. coli enzyme, and 0,3 rmole of pApApAp.4 (expressed us adcnosiric tinits). Incubation time, 1.5 hours at 37". The entire mixture was treated us de- scribed in the section on methods and the washed charcoal was eluted three times with 0.8 ml. of ethanolic ammonia. The eluate was concentrated, applied to paper, and chromstogruphed for 20 hours in System 2. The known compounds are on tahe left. (b) , ultraviolet photograph of paper electrophoresis strip demonstritting pApA 3s ii product of llie phosphorolysis of pApApA4pA. Area c of the chrornittogrurn shwn in (a) was eluted with water, concentri&ed, and aul-tjected to p:tper electro- phoresis (System 3) for 2 hours. phosphates isolntsed were devoid of any radioactivity. The phpA iso. li~ted from the phosphorolysis of pApApA was, in addition, treated with semen monoesterme as described for the preparation of the compounds of Group 11. The product hpA was identified by c€mmatogmphy in Solvent 1, where it had an Rsp-aMp (R, relative to the Rp of 5'-AMP) equal to 2.23. The Rs~-~~~~ of known ApA mas 2.23, and that for pApA was 0.42. TABLE IV Identification of ADP and UDP as Produets of Oligonucleotide Phosphorolysis pApApA.. . . .. . pApApApA.. . . . ApApA ... . . . .. . ApApApA.. . . . ApApU ... . . . . . . ApApApU. .. ... Substrate Concen- trationt 0.3 0.3 0.6 0.5 0.7 0.6 RF of JDP pro- duced' 0.42 0.36 0.48 0.30 0.17 0.22 0.45 Rr'-AMP authentic compound$ 0.40 0.44 1.9 2.0 2.0 2.0 NDP produced z;tff 0.45 0.43 1.9 2.0 2.2 2.2 Rp of uthentii NDP* 0.37 0.36 0.43 0.29 0.19 0.22 0.46 Mobility compound NDP eluted pmolc 0.024 0.014 0.041 0.064 0.110 0.108 0.045 Mobility found NDP ADP (6 64 I( UDP ADP I< c.9.m. peY pwlc 150,OOO 210,000 130,000 150,000 170, OOO 160,OoO 140,000 6.p.m. 3,790 2,m 5,400 9,700 18,400 16,800 6,500 180,000 190,000 140,OOO 150,OOO 160,000 160,000 160,OOO * Rp values in solvent System 2. t Specific radioactivity, in counts per minute per micromole, of inorganic P3* present in the incubation mixture. $ Substrate concentrations are expressed as nucleoside units present in a 0.15 ml. incubation mixture. These experiments were carried out in a manner similar to that described for the experiment of Fig. 3. The E. coli enzyme was used. The data for pApApApA here were obtained from that experiment. The nucleoside diphosphate area on the chromatogram was quantitatively eluted, its ultraviolet absorption determined to give the micromoles of ADP produced, and the radioactiv- ity of the eluate was determined on a suitable aliquot. Similar manipulations gave the data for the other compounds. TABLE V Identification of Dinucleotides and Dinucleoside Monophosphates as Products of Oligonucleotide Phosphorolysis System 2' System 1 System 3t Limit dinucleoside derivative Substrate RP authentic compound RF found 0.48 0.47 0.73 0.62 0.70 0.74 0.48 0.49 0.73 0.67 0.74 .O .74 9.1 10.9 4.7 8.9 10.9 4.6 4.0 1 4.4 * System used to separate dinucleoside derivative from nucleoside diphosphate t Mobility is calculated as cm. per 2 hours per lo00 volts applied. $ R6um gives the Rp relative to the Rp of 5'-AMP. The dinucleoside derivative end products were obtained from the same ehperiments described for Table IV, or from similar experiments. The experiment for pApApApA is described in detail in the text and in Fig. 3. The charcoal eluates were chromatographed in System 2 to separate the products, the pApA and ApA areas were eluted and concentrated, and samples were subjected to chromatography in System f and paper electrophoresis, 224 and nucleoside triphosphate. M. F. SINGER 225 In one experiment, all the products of phosphorolysis of pApApA were eluted quantitatively from a paper chromatogram run in System 2, and the yield of each substance was determined. It was assumed, for the purpose of calculation, that the pApA was contaminated by an amount of AMP equivalent to the ATP found, since both would have been formed in equal quantities from ADP by the action of myokinase. It was then possible to determine that the ratio of adenylic acid units recovered as pApA to those recovered as the sum of ATP, ADP, and AMP was 2.2. The theo- retical value given by Equation 2 is 2.0. DISCUSSION The data presented here describe the specificity of polynucleotide phos- phorylase with respect to the nature of the end groups on a polynucleotide substrate. It is clear that oligonucleotides which contain a monoesteri- fied phosphate at C-3' of the terminal nucleoside residue (ApApUp and ApApApUp) are resistant to phosphorolysis, and the same is true for the three oligonucleotides bearing a terminal 2`, 3'-cyclic phosphate. By con- trast, phosphorolysis occurs readily with comparable compounds which possess phosphate monoesterified at C-5` (pApApA and pApApApA) or contain no phospho monoester groups at all (for example, ApApA, ApApU). The resistance of relatively degraded commercial yeast RNA and RNA * "core" to phosphorolysis (3) is consistent with the present results, for these preparations are known to contain C-3' end groups (20, 24). Other factors such as molecular size must be considered, however, when compar- ing the rates of phosphorolysis of different RKA and biosynthetic polymer preparations. Ochocz has demonstrated that the state of aggregation of the polymer chains influences the rate of phosphorolysis (9). In the pres- ent work, comparisons have been made between oligonucleotides of the same chain length, so that the influence of end group structure could be studied without the complication of gross differences in molecular weight. A study of the products of phosphorolysis of oligonucleotides yields additional information concerning the mechanism of the polynucleotide phosphorylase reaction. The cleavage of the trinucleoside diphosphate, ApApA, can be used as an example. Phosphorolysis can occur in one of the two ways illustrated in Equations 3 and 4: ApApA + Pi 4 ApA + ADP ApApA + Pi --.) adenosine + ppApA (3) (4) Equation 3 involves cleavage at a nucleoside unit linked to the chain by its C-5'-hydroxyl, and the products are the pyrophosphorylated mono- nucleotide (ADP) and ApA. In Equation 4 phosphorolysis occurs at the 226 PHOSPHOROLYSIS OF OLIGONUCLEOTIDES nucleoside linked by its C3'-hydroxyl to the rest of the polynucleotide, and the products are adenosine and the pyrophosphorylated oligonucleotide, ppApA. In the experiments described above, the products of the phos- phorolysis of ApApA were ApA and ADP, indicating that the reaction proceeded according to Equation 3. Similarly, the phosphorolysis of ApApApU produced UDP and ADP, not adenosine and ADP, and the cleavage of the other oligonucleotides also proceeded according to Equation 3. This mechanism represents the reverse of a polymerization mechanism in which the 5'-nucleoside diphosphate units are added to the C3'-hy- droxyls of the acceptor nucleotides with the displacement of inorganic P from the mononucleotide. This phosphorolysis, according to Equation 3, is equivalent to the reverse of the polymerization Mechanism B discussed by Kornberg (25). Polymerization Mechanism A (25) represents the re- verse of Equation 4, and does not appear to be applicable to this enzyme, Equation 1, as written here and by others (2, 3, 6), implies that poly- nucleotide phosphorylase catalyzes the formation of polynucleotide chains from mononucleotides alone. Assuming this mechanism, it might be ex- pected that in the reverse direction, namely phosphorolysis, cleavage would result in the complete breakdown of a polynucleotide to mononucleotide units. The data presented above demonstrate that this is not the case. The phosphorolysis of oligonucleotides containing 3 or 4 nucleoside residues results in the accumulation of the compounds with 2 nucleoside units. In confirmation, Tables I and I1 show that pApA, ApA, and ApU are not phos- phorolyeed at significant rates. One possible explanation for these results may be that the enzyme demonstrates an exacting specificity when the substxate presented to it has only 2 nucleoside units. The condensation of 2 ADP molecules according to Equation 1 would be expected to form ppApA, a dinucleotide with a pyrophosphate end group. This compound has not yet been prepared, and it might undergo phosphorolysis. An alternative conclusion could be that polynucleotide phosphorylase, like starch phosphorylase (26) , catalyzes only a limited phosphorolysis of the chain, and the "limit oligon~cleotide'~ happens to be a dinucleotide. Recent experiments (27,28) have suggested that purified preparations of A. winehndii polynucleotide phosphorylase catalyze the condensation of 2 mononucleotide units to a dinucleotide very slowly, if at all. Thus, a lag in the polymerization reaction can be overcome by the addition of a pre- formed polynucleotide chain to the reaction mixture. In analogy with the enzymic synthesis of polysaccharides, the new polymer has been shown to be built onto this primer. It is consistent with the experiments reported here that the primer may be as small as a dinucleotide (28). M. F. SINGER 227 The author wishes to thank Dr. Leon A. Heppel for suggesting this problem and for many stimulating discussions during the course of the work. SUMMARY The phosphorolysis of oligonucleotides by preparations of polynucleotide phosphorylase from Azotobacter winehndii and Escherichia coli has been studied. Tri- and tetranucleotides with a phospho monoester group at the terminal C-5' were readily phosphorolyzed; however, if the phospho monoester group was at the terminal C-3', the compounds were resistant to enzymic attack. Trinucleoside diphosphates and tetranucleoside tri- phosphates were also phosphorolyzed by these enzymes, indicating that the C-5' phospho monoester moiety is not required in order that an oligonucleo- tide be a substrate. Dinucleotides and dinucleoside monophosphates were not phosphorolyzed at significant rates. The products of the phosphoro- lysis of the compounds with 3 and 4 nucleoside residues were identified as the nucleoside diphosphates (ADP or UDP) and the resistant dinucleotide or dinucleoside monophosphate. BIBLIOGRAPHY 1. Grunberg-Manago, M., and Ochoa, S., J. Am. Chem. Soc., 77, 3165 (1955). 2. Grunberg-Manago, M., Ortiz, P. J., and Ochoa, S., Biochim. el biophys. acta, 20, 3. Littauer, U. Z., andKornberg, A., J. Biol. Chem., 226, 1077 (1957). 4. Markham, R., and Smith, J. D., Biochem. J., 62, 558 (1952). 5. Grunberg-Manago, M., Ortiz, P. J., and Ochoa, S., Science, 122,907 (1955). 6. Ochoa, S., and Heppel, L. A., in McEIroy, W. D., and Glass, B., The chemical 7. Heppel, L. A., Ortiz, P. J., and Ochoa, S., J. Biol. Chem., 229,679 (1957). 8. Heppel, L. A., Ortiz, P. J., and Ochoa, S., J. Biol. Chem., 2!B, 695 (1957). 9. Ochoa, S., Arch. Biochem. and Biophys., 69, 119 (1957). 269 (1956). basis of heredity, Baltimore, 615 (1957). 10. Heppel, L. A., Ortiz, P. J., and Ochoa, S., Science, 123,415 (1956). 11. Singer, M. F., Federation Proc., 16, 250 (1957). 12. Ochoa, S., Special publication of the New York Academy of Sciences, New 13. Colowick, S. P., in Colowick, S. P., and Kaplan, N. O., Methods in enzymology, 14. Sevag, M. G., Lackman, D. B., and SmoIens, J., J. Biol. Chem., 124,425 (1938). 15. Beers, R. F., Jr., Biochem. J., 66, 686 (1957). 16. Markham, R., and Smith, J. D., Biochem. J., 62, 552 (1952). 17. Krebs, H. A., and Hems, R., Biochim. el biophys. acta, 12, 172 (1953). 18. Kornberg, A., and Pricer, W. E., Jr., J. Biol. Chern., 193,481 (1951). 19. Kornberg, A,, in McElroy, W. D., and Glass, B., Phosphorus metabolism, Balti- 20. Volkin, E., and Cohn, W. E., J. Biol. Chem., 206,767 (1953). 21. Privat de Garilhe, M., and Laskowski, M., J. Biol. Chem., 223, 661 (1956). York, 6,191 (1957). 2, 598 (1955). more, 1, 392 (1951). 228 PHOSPHOROLYSIS OF OLIGONUCLEOTIDES 22. Cook, W. H., Lipkin, D., and Markham, R., J. Am. Chem. SOC., 79,3607 (1957). 23. Sutherland, E. W., and Rall, T. W., J. Am. Chem. Soc., 79,3608 (1967). 24. Markham, R., and Smith, J. D., Biochem. J., 62, 565 (1952). 25. Kornberg, A., Advances in Enzymol., 18, 191 (1957). 26. Hestrin, S., J. Biol. Chem., 179, 943 (1949). 27, Mii, S., and Ochoa, S., Biochim. et biophgs. acta, 26,445 (1957). 28. Singer, M. F., Heppel, L. A., and Hilmoe, R. J., Biochim. et biophys. acta, 26,447 (1957).