BE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 235, No. 3, March 1960 Printed in U.S.A. The Polymerization of Guanosine Diphosphate by Polynucleotide Phosphorylase MAXIKE F. SINGER,* RUSSELL J. HILMOE, AND LEON A. HEPPEL born the National Institute of Arthritis and Metabolic Diseases, Natiod Institutes of Health, United States Public Health Service, Bethesda, Maryland (Received for publication, September 9, 1959) Polynucleotide phosphorylase catalyzes the reversible poly- tion of nucleoside diphosphates (1). With enzyme prep- ns from Escherichia coli (2), Azotobacter agile (3), and Mi- us lysodeikticus (4, 5) a reaction has been observed with additions of adenosine, uridine, cytosine, or inosine diphos- to form the corresponding homopolymer. However, the erization of guanosine diphosphate represents a special . Littauer and Kornberg (2) observed no polymerization tion upon incubating GDP with fractions from E. coli. In otide phosphorylase from A. agile, noted a very slow release of PI from ped far short of the equilibrium point . In contrast, when mixed with other leoside diphosphates, GDP is well utilized and polymers ly AGUC' are formed (3). eriments described below show that GDP, when pres- is not polymerized even after many hours by enzyme rom A. agile or E. coli. However, polymerization of does take place in the presence of oligonucleotides such as and ApApU, and guanosine monophosphate units are the primer. No reaction occurs with GDP and oligo- es, such as ApApUp, that do not contain an unsubsti- droxyl group at carbon 3' of the terminal nucleoside due. Highly purified fractions of A. agile polynucleotide phorylase, kindly supplied by Drs. Mii and Ochoa (6), cata- the polymerization of ADP, UDP, CDP, and IDP only af- a lag period. Thii lag can be overcome by ribonucleic acid ertain other polymers (6), as well as by oligoribonucleotides us types, including pApApA, ApApU and ApApUp (see g paper (7)). Accordingly, GDP differs from other nu- diphosphates in that its polymerization requires the sence of oligonucleotides of the type that can be incorporated MATERIALS AND METHODS t of the experimental procedures were the same as those preceding paper (7) and only brief reference will be made Research Fellow of the National Institute of Arthritis and abolic Diseases of the National Institutes of Health, United tes Public Health Service, 1956-1958. The abbreviations used are: EDTA, ethylenediaminetetra- tate. The biosynthetic polymers synthesized by the action olynucleotide phosphorylase are : poly A, polyadenylic acid ; y U, polyuridylic acid; poly C, polycytidylic acid; poly G, uanylic acid; poly AU, copolymer of adenylic and undylic oly AGUC, copolymer of adenylic, uridylic, guanylic, and c acids. RQDP is the ratio of the Rp of a given compound Materials--GDP was obtained from the Sigma Chemical Com- pany and samples were usually stored at 3" for several months before use. The preparations were found to contain up to 15% of 5`-GMP and up to 4% of GTP by quantitative paper chro- matography in Solvent 3 (see below) and the values stated for concentrations of GDP have been corrected. A sample of a chemically synthesized adenylic acid polymer (8) was kindly provided by Dr. A. M. Michelson, Arthur Guinness Son and Com- pany, Ltd., Chemist's Laboratory, Dublin; it consists of short polynucleotide chains containing mixed 3`-5` and 2`-5` phospho- diester bridges and terminated by a cyclic 2', 3'-phosphoryl group. A dinucleoside monophosphate with a 2`-5` phosphodiester bridge, namely adenylyl-(2'-5')-uridine (9) was also a gift from Dr. Michelson. The preparation of polynucleotide phosphorylase from E. coli was a fraction carried through the first ethanol step in the pro- cedure of Littauer and Kornberg (2); its specific activity in the "exchange" assay (3) was 15. Three highly purified prepara- tions from A. agile (6) were given to us through the kindness of Drs. Mii and Ochoa; all of them were found to display a lag period in the polymerization of ADP and UDP which could be overcome by addition of a suitable primer. The fractions are designated by their specific activity in the exchange assay (3) as S.A. 70, S.A. 60, and S.A. 350. Two fractions (S.A. 40 and S.A. 160) that did not show well defined lag periods were also obtained from Drs. Mii and Ochoa. Snake venom phosphodiesterase was prepared by a modification of the procedure of Koerner and Sinsheimer (10). Phosphomonoesterase was fractionated from human seminal plasma (1 1). MeUlook-The solvent system used for descending chroma- tography were: Solvent 1, isopropanol-water (70:30, volume for volume) with NHs in the vapor phase (12); Solvent 2, saturated ammonium sulfate-isopropanol-1 M sodium acetate (80 : 2 : 18, volume for volume for volume) (13); Solvent 3, isobutyric acid-1 M NH40H-0.2 Y EDTA' (100:60:0.8, volume for volume for volume) (14); Solvent 4, isopropanol, 170 ml, concentrated HCl, 44 ml, water up to 250 ml (15); Solvent 5, n-propanol-concen- trated NH40H-water (60:30:10, volume for volume for volume) (16). Whatman No. 3MM paper was used with Solvents 1, 3, and 5, and Whatman No. 1 paper with Solvents 2 and 4. Inorganic phosphate was determined by the method of Fiske and SubbaRow (17), with the use of the Klett colorimeter with a No. 66 filter. The higher concentrations of oligonucleotide used in this study caused an interfering turbidity upon addition of am- monium molybdate. This difficulty was overcome and satis- factory assay of Pi was obtained by the following procedure. An aliquot of the incubation mixture containing from 0.015 to 0.060 751 Polymerization of GDP Vol. 235, No. 3 TIME IN HOURS FIG. 1. Polymerization of GDP as measured by formation of Pi; effect of primers. The incubation mixture (0.05 ml) contained 328 pg per ml of enzyme for Curve 3 and 164 rg per ml for all of the other curves. The concentration of GDP was 6.8 mM for Curves 1, 8, 3, and 13.7 mM for Curves 4, 6, and 6. Primer addi- tions: Curve 1, 4.6 mM pApApA; Curve 8, 3.8 mM pApA; Curve 3, 1.3 rn pApA; Curve 4, 1.8 my pApApA; Curve 6, 3.8 mM pApA; Curve 6, 7.1 mu pApA. Aliquots of 0.01 ml were removed for Pi analysis. In control experiments, without primer, the concen- tration of Pi was <0.1 mM after as long as '20 hours. 5 i I fY4 w a .- 3 a v, wJ2 0 I *I I I 0.006 M GDP PRIMER =0.004 M pApA 164 I 2 U 0 TI ME (HOURS) FIG. 2. Polymerization of GDP with different amounts of en- zyme. The incubation mixtures contained the amounts of enzyme shown in the figure (pg per ml), and the following in mM: GDP, 6; pApA, 4.0. pmole of Pi is made up to 0.5 ml with cold perchloric acid whose strength is such that the final concentration is 2.5%. After 5 minutes at 0" the mixture is centrifuged and 0.3 ml of the super- natant fluid is mixed with 0.8 ml of 1 N H$Oa, 0.26 ml of water, 0.16 ml of 2.5% ammonium molybdate, and 0.08 ml of reducing reagent. The mixture is centrifuged for 5 minutes at 1,500 X g in the International No. 1 centrifuge. The clear, blue, super- natant fluid is decanted carefully into a Klett tube and read in the instrument, together with appropriate phosphate standards and blanks, 10 minutes after addition of the reducing agent. EXPERIMENTAL AND RESULTS E&ct of pApA and its Homologues on Polymerization of GDP- There is no polymerization reaction when GDP alone is incubated with polynucleotide phosphorylase from A. agile. Thus, no re- lease of Pi (<0.02 pmole) from GDP occurs in 4 hours, with amounts of enzyme which would form 10 to 20 pmoles of Pi from ADP per hour.* Over 30 experiments were carried out with 6 different lots of enzyme, with 0.005 to 0.05 M GDP and incubation times that varied from 4 to 24 hours. In addition, study of the reaction mixture by paper chromatography in Sol- vents 1, 2, and 3 affords no evidence for the formation of poly- nucleotide material. However, in the presence of pApA, pA- pApA, or pApApApA one observes formation of Pi at a rate that is nearly linear with time until equilibrium is approached. The equilibrium point is not well defined by the data but it ap- pears to correspond to the conversion of from 70 to 80% of GDP to Pi and polynucleotide (Fig. 1). It can be seen from Fig. 2 that the initial rate of formation of Pi is proportional to the concentration of enzyme, in the pres- ence of 6 mM GDP and 4 mM pApA. A similar rate was ob- tained at 13.7 and 30 m~ GDP, suggesting that the enzyme is saturated with GDP at a concentration of less than 6 nm. Be- cause of insufficient material it was not possible to determine the concentrations of the various primers required to saturate the enzyme. Thus the rate of Pi formation is 0.9 and 2.0 pmoles per hour per ml of reaction mixture in the presence of 4.3 and 7.0 mM pApA, respectively (GDP, 13.7 mM; enzyme, S.A. 70, 82 pg per ml). The trinucleotide is effective at lower concen- trations; under the same conditions mentioned above 2.8 mM pApApA gives a rate of 3 pmoles of Pi per hour per ml. It should be noted that concentrations of 'PAPA and pApApA that do not saturate the GDP system are more than enough to give maximal stimulation of the rate of polymerization of ADP and UDP (7). The tetranucleotide, pApApApA, was tested at con- centrations of 2.1 mM, 0.74 mM, and 0.44 mM; these levels were found to be approximately equivalent to 2.9 mM, 0.97 mM, and 0.58 mM pApApA, respectively. Thus, the tetranucleotide is effective at somewhat lower concentrations than required for the trinucleotide. There is no detectable formation of Pi when any of the enzyme fractions are incubated with pApA, pApApA, or pApApApA in the absence of GDP. Also, there is no reaction in the absence of enzyme. Similar results were obtained with the E. coli polynucleotide phosphorylase. Again, in the absence of pApA or pApApA, no material with an Rp of zero was detected on paper chroma- tograms nor were any oligonucleotides containing guanosine formed. The curve of Pi formation plotted against time was also quite fiat except for a small initial burst. A minor con- taminant of GDP preparations appears to be rapidly dephos- phorylated by the E, coli enzyme but not by A. agile fractions. It might also be mentioned that a primer requirement for GDP polymerization could be demonstrated with cruder enzyme prep- * In the case of enzyme fractions that showed a lag phase in the polymerization of ADP this comparison is based on the rate achieved after the lag phase, or on the initial rate in the presence of saturating amounts of poly A. rations froin A. ugite , preparatiom that sl~oowod no primer re- uiremrnt for the polymerization of ADP and UDP. .itft&mbLq That Were Tested rn Primers-Table kt of nucleotides that did not induce polymerization In addition, no effect was noted with an unknown ma- wcum in filter pper and can be eluted with water, experiments with ADP or WDP a small hut signifi- oligonucleotides containing no phoaphomonoeater end ere tested with 13.7 mw GDP and 82 pg of S.A. 70 en- r rnl. Thc results were HS follows: (a) ApA, at 3.6 m~, in a stirnulittion of P, release equivalent to that ob- h 4.2 mhl pApA, (b) ApApU (2 m~) had exactly the its pApApA (2 mar). ApApA was less deftivc, but nre of tlasc (SA. TO) wus u8ed. Except where specifically indicaled, the ~:L-ID~ iiirubtrtiort con- ditiaris and Lhe same enzyme preparation were used in the other cxperirnent8 reported. _____ ___~ _lll~" - Compound 1 Concentration msh1 1.2 1.0 3.0 3.0 1.2 N. F. Singer, R. J. Hilmoe, and L. A. Iieppel 753 1--------- - _" , .." ., 1 ,I AD~~P ApAplJp Adeaosirie AMP Adenosine 5'-henzyl phosphate Adenylyl (2'-5')-uridine maI 5.x 0.8 5.0 6.6 5.0 6.3 FJG. 3. Tntmviolet photograph of x chromatogram run in 801- vent 3. At zero lime (riot shown) the only visible densities corresponded tu pAyhpA, GMP, GDP and GTP. This photograph shows partial disappearwee of pApApA as the incubation proceedli, as well as the formation of pApApApG, p ApApApGpG, pApApApCpGpC , and polymer (visible at 63 minutes, near top of photograph). Tlk3 is from an experiment similar to that. in Pig. 4. in 1 N RCI (18) follou~ecl by quantitative chromatography in Sol- vent 4 yieldcd denine and guanine in a ratio of 3.2: 1.0 (theory, 3.0:l.O). Evdrolvsis in 0.3 N KOH (19). followed bv chroma- 754 Polymerization of GDP Vol. 235, No. 3 TIME IN HOURS FIG. 4. Polymerization of GDP in the presence of pApApA. The incubation mixture (0.2 ml) contained 82 pg per ml of enzyme. and the following, in pmoles per ml: GDP, 13.7 and pApApA, 2. Aliquots of 0.03 ml were chromatographed in Solvent 3 and the various bands were quantitatively transferred and run in Solvent 1. Quantitative elution was again carried out and the concentra- tion of the various nucleotides was determined.J No polymer could be detected after 15 minutes but it was found in samples removed after 40 minutes of incubation. Thereafter the concen- tration of polymer increased progressively; it reached a value equivalent to 8 pmoles of base per ml in 6.5 hours, when it ac- counted for almost all of the GDP which disappeared. tography in Solvent 1 gave adenosine 3`, 5'-diphosphate (mixed with 2',5` isomer), 3'-AMP (mixed with 2`-AMP), and guano- sine, the expected products. No guanylic acid was formed. The results of partial hydrolysis by purified snake venom phos- phodiesterase were also informative. This fraction has been shown to cleave both oligodeoxyribonucleotides (20) and poly- ribonu:ieotides (21) in stepwise fashion, beginning at that end of the molecule bearing an unsubstituted hydroxyl group at car- bon 3'. From pApApApG one would expect 5'-GMP and pA- pApA to be the major early products, followed later by pApA and 5'-AMP. When enzymic hydrolysis of pApApApG was carried to the extent of 40% the following were found by quantitative elution from paper chromatograms: pApApA, 0.023 pmole; 5`- GMP, 0.020 pmole; 5'-AMP, 0.010 pmole; pApA, 0.002 pmole. Evidence for the structure of pApApApGpG is the following: It moves as a single band in Solvents 1 and 3, with a lower RF than the tetranucleotide just discussed. Acid hydrolysis yielded adenine and guanine in a ratio of 1.7 : 1.0 (theory, 1.5: 1 .O). Al- kaline hydrolysis gave adenosine 3`, 5`-diphosphate (and the 2', 5' isomer), 3`-AMP (and 2'-4MP), 3'-GMP (and 2'-GMP), and guanosine. The hexanucleotide, pXpXpApGpGpG, was obtained in small amounts. It was found to contain adenylic and guanylic acid residues but there was insufficient material for accurate quanti- AS in the preceding paper (7), the assumption is made that the molar extinction coefficient of an oligonucleotide is approxi- mated by the sum of the extinction coefficients of its constituent nucleotides. Values used for the molar extinction coefficients, at 257 mp and pH 2, for AMP and GMP are 15,100 and 12,200, re- spectively. The approximation does not account for any hypo- chromic effect. tative analysis. The structure is therefore only tentatively as- signed. The following values for RoDP were obtained for pApApApG, pApApApGpG and pApApApGpGpG: in Solvent 1, 0.52, 0.26, and 0.10, respectively; in Solvent 3, 1.60, 1.17, and 0.49, re- spectively. The ratio of absorbancy (pH 2) at 280 mp over that at 260 mp was found to be 0.33 for pApApApG, 0.42 for pApA- pApGpG, and 0.58 for pApApApGpGpG. Structure of "poly G"-This term is restricted to material that is precipitated by 2 volumes of ethanol, that is insoluble in cold 2.5% HC104, and that remains at the origin upon chromatog- raphy in Solvents 1, 2, or 3. Alkaline hydrolysis of such ma- terial yields mostly 3`-GMP and 2'-GMP, identified by their Rp values in Solvents 1 and 2, and by the fact that they were read- ily hydrolyzed by phosphomonoesterase purified from human seminal plasma but not by 5'-nucleotidase. Smaller amounts of adenine-containing nucleotides derived from the incorporated primer are also obtained. The following experiment illustrates the preparation of "poly G," of short average chain length, formed by addition of guano- sine monophosphate units to pApApA. The incubation mixture contained 50 pg of E. coli polynucleotide phosphorylase, 125 pmoles of Tris buffer, pH 8.2, 10 pmoles of MgC12,0.4 pmole of EDTA, 1.31 pmoles of pApApA, and 27.4 pmoles of GDP, in a total volume of 1.0 ml. The mixture was incubated at 37" for 25.5 hours, with toluene added after 6 hours. The formation of Pi amounted to 8 pmoles. A chromatogram run in Solvent 3 showed complete incorporation of pApApA and all of the reac- tion products had an Rp of zero. Two volumes of cold ethanol were added and after 3 hours at 2" the precipitate was collected by centrifugation and dissolved in distilled water. The solu- tion was dialyzed against 1 liter of cold 0.001 M EDTA for 24 hours and then against cold, running distilled water for 48 hours. The yield of polymer, based on measurement of optical density, was 56% of what could be expected from the amount of Pi that had been formed. Paper chromatographic examination of the product after dialysis showed complete removal of GDP. Hy- drolysis of a sample of the dialysis residue in 1 N HC1 (18) fol- lowed by chromatography in Solvent 4 gave a guanine-adenine ratio of 2.8: 1, corresponding to an average chain length of 11.4. This indicates the addition of 8.4 guanylic acid residues, on the average, to every molecule of pApApA. The material was acid insoluble, and was nondialyzable. In later esperiments, with the use of much larger amounts of enzyme, poly G that had an average chain length of 30 was syn- thesized. Further, with the use of pUpUpU as a primer and treating the resultant polymer with pancreatic ribonuclease, it was possible to remove uridylic acid from the preparation and SO obtain poly G free of bases other than guanine. In a typical esperiment the incubation mixture (0.5 ml) contained 27.4 pmoles of GDP, 75 pmoles of Tris buffer, pH 8.2, 5 pmoles of MgC12, 0.2 pmole of EDTA, 0.9 pmole of pUpUpU, and 320 pg of a gel eluate fraction from A. agile (S.A. 160). The mixture was kept at 37" for 6.5 hours. The polymer was precipitated by the addi- tion of 1.0 ml of cold 50/, HClOa and collected by centrifugation. It was washed with 2 ml of 3% HCIOl and then with two por- tions (2 ml each) of 0.01 N HC1. The precipitate was suspended in 1 ml of water and dissolved by addition of sufficient 1 N NH40H to bring the pH to 7.0. An opalescent, distinctly vis- cous solution was obtained. An aliquot was hydrolyzed with 0.3 N KOH (19) and the products were separated in Solvent 5 and March 1960 M. F. Singer, R. J. Halmoe, and L. A. Heppel 755 0- 6.8 X M GDP - I I. u, 6.8 X M GDP t 4. I x 10-3~ PAPAPA u, FIQ. 5. Polymerization of GDP in the presence of pApApA. 'he incubation mixture (0.2 ml) contained 82 pg per ml of enzyme nd the following, in pmoles per ml: GDP, 6.8; pApApA, 4.1. lamples of 0.03 ml were removed at intervals for chromatography 2 Solvent 3. In B, hanges in the concentration of GDP and various oligonucleotides uantitatively eluted. The products were: uridine 3', 5'- (and ', 5')-diphosphate, 0.05 *mole; 3' (and 2')-GMP, 1.28 pmoles; ' (and 2')-UMP, 0.105 pmole; guanosine, 0.045 pmole. From he total amount of compounds containing guanosine isolated in his esperiment the yield of poly G was calculated to be 50% of hat expected from Pi release. The results are close to what one iould expect from a polymer with an average chain length of 0. This can be represented diagrammatically, with vertical .otted lines showing the points of alkaline cleavage : In A, formation of P, plotted against time. ...... ...... a a z w -J w 0 I 2 3 4 TIME IN HOURS - -0 I 2 3 4 TIME IN HOURS plotted against time (Curve A, GDP; Curve B, pApApA; Curve c, pApApApG; Curve D, pApApApGpG; Curve E, pApApApGpGpG). Two other products were found, but only in the sample removed at 240 minutes, namely pApA (0.35 mM) and polymer (equivalent to 2.0 mM mononucleotide). Time Course of Polymerization Reactwns with Diflerent Con- centrations of GDP and Primer-Fig. 4 shows the results of an experiment in which GDP and pApApA were incubated with the A. agile enzyme. At various time intervals, aliquots of the in- cubation mixture were removed and its components were deter- mined by quantitative chromatography. It is evident from Fig. 4 that most of the primer is utilized within 15 minutes and all of it disappears within several hours. The concentration of pApA- pApG rises to nearly 0.7 mM in 40 minutes and after 6 hours falls to 0.1 mM. The pentanucleotide, pApApApGpG, accumulates only in trace amounts (not shown), whereas the hexanucleotide reaches a level of 0.11 mM in 4.5 hours, No polymer is noted after 15 minutes but it can be found in an aliquot removed after 40 minutes and thereafter its concentration rises (see legend to Fig. 4). An experiment with 13.7 mM GDP and 1 mM pApApA gave similar results. An experiment was carried out under identical conditions ex- cept that the concentration of pApApA was increased to 4 mM. This led to a more rapid accumulation of pApApApG, so that its concentration was 1.6 mM after 20 minutes. Also there was a less rapid decline with time, compared with the experiment men- tioned above. A substantial amount of pApApApGpG was formed, as well as smaller amounts of pApApApGpGpG. KO polymer was evident in a sample removed after 20 minutes but it was found after 1 hour, and thereafter increased with time of incubation. In this experiment pApAp-4 was not completely utilized; in fact, its concentration showed a small rise between 1 and 6 hours. The results of an experiment in which the concentration of GDP (6.8 mM) was brought close to that of pApAp-4 (4.1 mM) are presented in Figs. 5A and B. The changes to be observed between 40 and 240 minutes illustrate several points. First, there is a substantial decrease in the concentration of pApA- pApG. Also, there is evidence of a continued forward reaction, Possible explanations are considered below. t Polymerization of GDP Vol. 235, KO. 3 e in GDP (-0.9 m~), (b) increase in Pi ), and (c) appearance of polymer (equivalent to 2.0 . There is also evidence for a net reaction in the re- Thus, no pApA is after 20 and 40 minutes, but 0.35 my4 is noted after 240 This simultaneous phosphorolysis of pApApA and polymerization may esplain why the decrease in GDP than the net increase in P,. similar experiments, except that more enzyme was present, s found that the composition of the reaction mixture changed able estent after the formation of Pi and the utiliza- came to a halt. Disappearance of pApApApG and pG was noted, together with an increase in the con- ion of polymer and of smaller oligonucleotides. These us to incubate pApApA and pApApApA with osphorylase in the absence of added nucleoside i. A transnucleotidation reaction was found multaneous formation of smaller and larger ynucleotides. This reaction does not appear to involve the rticipation of ADP or Pi, at least not in the free state. A de- study will be reported in a future publication. incorporation of the dinucleotide, pApA, into poly G was s estensively investigated, but here too it was found that oligo- cleotides containing guanine accumulated early in the course the reaction. At later time intervals these were observed to rease in concentration and material with an RP of zero \vas ection, occurring simultaneously. DISCUSSION The present results help to explain why no polymerization ction has been observed with GDP and polynucleotide phos- orylase in the past, even though guanosine monophosphate its have been incorporated into polynucleotide chains when a xture of nucleoside diphosphates was used. The data indicate t a polynucleotide cannot be synthesized de wuo from GDP; is only possible for the enzyme to catalyze the addition of ne monophosphate units to a preformed oligonucleotide These results were obtained with every sample of poly- ide phosphorylase that was tested. Some of these en- ons showed no primer requirement at all with the diphosphates. Thus, it should be emphasized s differ from those reported in the preceding a lag period in the polymerization of ADP or vercome by compounds such as pApApA or ApApUp. GDP the requirement for an oligonucleotide of suitable re appears to be absolute; without it no reaction can be d even after many hours. primer must have an unsubstituted hydrosyl group at on 3' of the terminal nucleoside residue in order for a 3'-5' odiester bridge to be established. Thus, ApUp and Ap- Compounds that contained adenosine or the terminal nucleoside residue were used in the pres- er for the polymerization of GDP must have least one 3`,5' phosphodiester bond. It is of considerable erest that ApA is an effective primer, whereas adenylyL(2'- uridine (which contains a 2'-5' phosphodiester bond) is not. was not possible to obtain an accurate measure of the ADP d by phosphorolysis along with pApA because it appeared ery diffuse densify on the chromatogram. inactive. No reaction could be demonstrated with poly A, poly U, poly C, or poly AGUC. Here there may have been inhibitory interac- tions, similar to the suppression of poly A synthesis by poly U (6). It is also possible that the additions of polymer that were employed provided too low a concentration of terminal nucleoside residues with free hydroxyl groups at C-3'. An effort was made to detect some reaction in an incubation mixture containing 10 mg per ml of poly A. Relatively large aliquots were removed for estimation of Pi, but the results were inconclusive. It is proba- ble that a primer must be a ribose derivative since pTpT and pTpTpT were inactive. In the presence of 4 mM pApApA the rate of Pi formation from GDP with purified A. agile fractions was one-fourth of that ob- tained with ADP. It is likely that even faster rates could have been observed with higher concentrations of primer. In other studies6 it was shown that under suitable conditions the rate of the GDP-Pi exchange reaction is nearly equivalent to that observed with other nucleoside diphosphates. It therefore appears that GDP is an effective substrate for the enzyme. These observa- tions are consistent with the findings of Mii and Ochoa (6), with highly purified fractions from A. agile, that indicate that a single enzyme is active with all of the ribonucleoside diphosphates. At present one may only speculate as to why there is an abso- lute requirement for a primer in the polymerization of GDP by polynucleotide phosphorylase whereas the same preparations of enzyme, when tested with other nucleoside diphosphates, may show only a lag period, after which a rapid reaction occurs in the absence of added primer. SUMMARY No polymerization of guanosine diphosphate occurs when poly- nucleotide phosphorylase is incubated with this compound alone. However, a polymerization reaction does take place in the pres- ence of an oligoribonucleotide containing an unsubstituted hy- droxyl group at carbon 3' of the terminal nucleoside residue. Such oligoribonucleotides include ApA, ApU, ApApU, pApA, and pApApA. The oligonucleotide serves as a primer and successive guano- sine monophosphate units are added to it, beginning with esteri- fication of the hydrosyl group at carbon 3'. Thus, with a trinu- cleotide containing adenosine serving as primer, polynucleotides were recovered whose structures corresponded to the addition of 1, 2, and 3 guanosine monophosphate residues. A polymer that is precipitated by acid and by 2 volumes of ethanol is also formed. It is nondialyzable and its composition indicates that, on the average, up to 27 guanosine monophosphate residues have been added to each molecule of primer. REFERENCES 1. GRUNBERG-MANAGO, M., AND OcnoA, S., J. Am. Chem. soc., 77, 3165 (1955). 2. LITTAUER, U. Z., AND KORNBERG, A,, J. Biol. Chern., 226, 1Oii (1957). 3. GR~NBERG-MANAA, M., oRTIZ, P. J., AND oCH~.~, s., Bio- chim. et Biophys. Acta, 20, 269 (1956). 4. BEERS, R. F., JR., Biochem. J., 66. 686 (1957). 5. OLMSTED, P. S., Biochim. et Biophys. Acta, 27, 222 (1958). 6. Mrr, S., AND OCHOA, S., Biochim. et Biophys. Acts, 26, 445 7. SINQER, M. F., HEPPEL, L. A., AND HILMOE, R. J., J. Biol. (1957). Chem., 236, 738 (1960). 5M. F. Singer, M. Grunherg-Manago, and R. J. Hilmoe, in preparation. IMarch 1960 M. F. Siwe~, R. J. Hilmoe, and L. A. Heppel 757 8. MICHELSON, A.M., Nature, 181, 303 (1958). 9. MICHELSON, A. M., SZABO, L., AND TODD, A. R., J. Chem. 10. KOERNER, J. F., AND SINSHEIMER, R. L., J. Biol. Chem., 228, 11. MARKHAM, R., AND SMITH, J. D., Biochem. J., 62, 558 (1952). 12. MABKHAM, R., AND SMITFI, J. D., Biochem. J., 62, 552 (1952). 13. MARKHAM, R., AND SMITEK, J. D., Biochem. J., 49,401 (1951). 14. KREBS, H. A., AND HEMS, R., Biochim. et Biophys. Acta, 12, 15. WYATT, G. R., Biochem. J., 48, 584 (1951). Soc., 1646 (1956). 1049 (1957). 172 (1953). 16. BANES, C. S., AND ISHERWOOD, F. A., Nature, 164,1107 (1949). 17. FISKE, C. H., AND SUBBAROW, Y., J. Bio2. Chem., 66, 375 18. SMITE, J. D., AND MARKHAM, R., Biochem. J., 46, 509 (1950). 19. CROSBIE, G. W., SMELLIE, R. M. S., AND DAVIDSON, J. N., 20. RAZZELL, W. E., AND KHORANA, H. G., J. Am. Chem. Soc., 21. HILMOE, R. J., Ann. N. Y. Acad. Sci., 81,660 (1959). 22. SEVAQ, M. G., LACKMAN, D. B., AND SMOLENS, J., J. Biol. (1925). BiocLm. J., 64, 287 (1953). 80, 1770 (1958). Chem., 124, 425 (1938).