THE MECHANISM OF ACTION OF POLYNUCLEOTIDE PHOSPHORYLASE L. A. Heppel, M. F. Singer, R. J. Hilmoe National InnstitGte of Arthritis and Metabolic Diseases, Public Health Service, Bdhesda, Md. The purpose of this paper is to review certain aspects of the mechanism of action of polynucleotide phosphorylase and to present, rather briefly, some recent findings. The discussion will be concerned with studies carried out by S. Ochoa and his associates at New York University, New York, N. Y., and with work done at the National Institutes of Health. Reference will also be made to some recent work carried out in Bethesda by Grunberg-Manago. Some of the material to be presented has already been published, but a review may be profitable at this time. Certain of the unsolved problems that face investigators in this field also will be discussed. Polynucleotide phosphorylase was discovered by Grunberg-Manago and Ochoa in extracts of Azotobacter agile.'r2 Studies of the nature of nucleotide incorporation into nucleic acid in Escherichia coli led to a recognition of the same reaction by Littauer and Rornberg.3, 4, Beers6 has made extensive studies of the enzyme from Micrococcus lysodeikticus, and some of this work will be presented in another paper in this symposium. Olmsted' has also reported studies dealing with polynucleotide phosphorylase from M. lyso- deikticus. The reaction catalyzed by the enzyme may be formulated as follows: (1) Mg++ 12 nucleoside-pp + (nucleoside-P)" + n Pi where Pi is inorganic phosphate and nucleoside-pp represents a nucleoside 5'-diphosphate. Polymers are formed in this reaction that have all of the struc- tural features of isolated ribonucleic acid (RNA) preparations and are attacked in a similar way by hydrolytic enzymes; the experimental evidence will not be reviewed here. In the forward direction, one can measure the release of inorganic phosphate or the formation of acid-insol- uble polymer; in the reverse direction, one can assay the rate of phosphorolysis of polymers such as RNA or adenylate polynucleotide (poly A), or of suitable oligonucleotides. Finally, one can measure the P!' nucleoside diphosphate- exchange reaction. In this assay, adenosine diphosphate (ADP) or other nu- cleoside diphosphate is incubated with enzyme- and P32-labeled inorganic phos- phate, and the ratio of these components is such that no detectable net forward reaction occurs; the amount of radioactivity incorporated into ADP is then determined. It is of interest to see how these different assays compare quantitatively when each is performed under optimum conditions. TABLE 1 shows data for an E. coli fraction (first ethanol step) prepared by Hilmoe according to Littauer and Kornberg15 and for a fraction from A. agil8 supplied by S. Mii and S. Ochoa. The results for all of these assays are given in the same units-that is, pmoles The reaction can be followed in various ways. 635 636 Annals New York Academy of Sciences per hour per milligram of protein. It is evident that measurement of Pi release in the forward reaction (EQUATION 1) gives much higher values for specific ac- tivity of the enzyme fractions than is true for the other assays. Some pos- sible explanations for these differences will become obvious as each of the three activities of the enzyme is discussed in turn. The forward reaction, which will be considered first, is most commonly fol- lowed by measurement of the rate of Pi formation from a single nucleoside diphosphate, or from a mixture of nucleoside diphosphates. With the enzyme preparation from E. coli or the earlier fractions obtained during purscation of the Azotobacter enzyme, polymerization occurs at a linear rate until equilib- rium is approached. Further, there is no stimulation, or priming, by the addi- tion of polymers or oligonucleotides (this is in contrast to results with highly purified Azotobacter fraction, described below). It should be emphasized, how- ever, that oligonucleotides of suitable structure may participate in the polymeri- zation reaction even in cases where they afford no stimulation of enzyme activ- ity; they are, in fact, incorporated into the polymer that is formed. TABLE 1 COMPARISON OF THREE ASSAYS FOR POLYNUCLEOTIDE PHOSPHORYLASE Micromoles per hour per milligram of protein E. coli I A. agile Pi formation from ADP* Phosphorolysis of poly At ADP-Pf2 exchange1 280 20 35 1000 45 100 * Assay described in Mii and Ochoa.8 t Assay described in Singer.I2 1 Assay described in Grunberg-Manago et aL2 An example of the type o€ oligonucleotide incorporated is the trinucleotide, pApApA.* This compound contains a phosphomonoester group at carbon 5' of the first adenosine residue, and the terminal nucleoside contains an unsub- stituted hydroxyl group at carbon 3'. Polynucleotide phosphorylase catalyzes the addition of a mononucleotide unit to the terminal nucleoside of pApApA, and this process continues, giving a polymer chain with pApApA forming its beginning portion. A considerable advance in our understanding of polynucleotide phosphoryl- ase came when Mii and Ochoas discovered a lag phase in the polymerization of ADP, inosine diphosphate (IDP), uridine diphosphate (UDP), and cytidine diphosphate (CDP), which was overcome by addition of RNA or of certain of the biosynthetic polymers. This lag period was found only with highly purified- A. agile fractions. Singer et aL9 found that oligonucleotides such as pApApA also overcame the lag peri0d.t In addition to stimulating the reaction in this * This and other abbreviations used follow the system described in the "Instructions to Authors" in The Journal of Biological Chemistry, September, 1958. t These studies were carried out with the A. agile fractions provided by Ochoa. Recently, the essential findings were confirmed with a fraction purified from A. agile by Singer and Hilmoe. However, the lag period was not as striking as in the earlier work and further purifi- cation of their fraction is indicated. Heppel et al. : Action of Polynucleotide Phosphorylase 637 manner, they were incorporated into the polymer by the same mechanism out- lined above for more crude Azotobacter fractions. These studies were extended to include oligonucleotides in which the hydroxyl group at C3` of the terminal nucleoside residue was blocked by a phosphomono- ester residue. With this compound (and its homo- logues) esterification to add new mononucleotide units and, thereby, to lengthen the chain is impossible. Consequently, oligonucleotides of this type are not incorporated, yet the surprising observation was made that they stimulate the An example is ApApUp. 0 W cn w _I w K a d- cn w HOURS FIGURE 1. The effect of ApApUp on the lag in the polymerization of ADP and UDP with purified A. agile polynucleotide phosphorylase. The reaction mixtures contained Tris buffer (PH 8.2), 7.5 pmoles; ethylenediamine tetraacetate, 0.02 pmole; ADP or UDP, 3 pmoles in a total volume of 0.05 ml. In the ADP experiments the reaction mixtures also contained 0.5 pmole MgClz and 5.1 X mg. enzyme (specsc activity 150 by the "exchange" assay); in the UDP experiments they contained 1.5 pmoles MgCL and 1.1 X mg. enzyme. The concentrations of oligonucleotides are indicated on the figure. The reaction was followed by determination of the release of P,. Incubation temperature, 37' C. polymerization reaction, overcoming the lag period both for ADP and UDP (FIGURE 1). There are no data to explain how these nonincorporated oligonucleotides act in overcoming the lag period. It is possible that such compounds and the various polymers stimulate the reaction in a similar fashion. However, there is no apparent specificity to be observed with oligonucleotides, while specificity relationships have been found with polymers: The great interest in primers that are not incorporated into new chains lies in the possibility that large poly- nucleotide molecules of this kind may have a directing influence on the compo- sition of the polymer synthesized. 63 8 Annals New York Academy of Sciences The behavior of guanosine diphosphate (GDP) is unique because, when pres- ent alone, it cannot be polymerized by any available preparation of polynucleo- tide phosphorylase, whether from E. coli or A. agile. There is no reaction with a large excess of enzyme or after many hours of incubation, even with those fractions that show no lag period with other nucleoside diphosphates. With the same enzyme preparations, GDP is well utilized if mixed with other nucleo- side diphosphates; thus, Grunberg-Manago et a1.2 described the preparation of poly AGUC several years ago. 6- i z 0.002 M pApApA - a - 0 2 4 6 TIME (HOURS) FIGURE 2. The polymerization of GDP in the presence of pApApA. The reaction mix- tures contained Tris buffer ($23 8.2), 23 pmoles; MgClz , 1.5 pmoles; ethylenediamine tetra- acetate, 0.06 pmole; GDP, 2.1 pmoles; and Azotobacler polynucleotide phosphorylase, 0.012 mg., in a ha1 volume of 0.15 ml. The experiment shown in the upper curve included, in addition, 0.3 pmole pApApA. The reaction was followed by determining the release of in- organic phosphate. Incubation temperature, 37' C. If an oligonucleotide with a free C-3' hydroxyl group is included in the incubation mixture, a polymerization reaction involving GDP does take place (FIGURE 2). An example of such an oligonucleotide is pApApA, but others would also serve. forming a phosphodiester bond and adding the first guanosine monophosphate residue: The hydroxyl group is estersed in the enzymatic reaction,, The tetranucleotide, pApApApG, is the first major product of the reaction; it has been separated by paper chromatography followed by rechromatography Heppel et al. : Action of Polynucleotide Phosphorylase 639 in another solvent system. Hydrolysis in 1N HC1 yielded adenine and guanine in a ratio of 3.2: 1.0, the theoretical being 3.0: 1.0. Digestion by alkali gave the expected products, adenosine 3`, 5'-diphosphate, adenosine 3'-phosphate, the corresponding 2'-isomers, and guanosine. Partial hydrolysis with snake venom phosphodiesterase (see Hilmoe, this monograph) gave the expected prod- ucts. The primer, pApApA, disappears as it is incorporated. The concentration of the compound just discussed, pApApApG, first rises and then falls as the addition of guanosine FIGURE 3 illustrates the time course of the reaction. I I I .4k -I* I .2 -/PAPAPA -12 I .o -- - 6.0 m D cn m 'F 0 r m cn 0 2 4 6 HOURS FIGURE 3. The polymerization of GDP in the presence of pApApA. The incubation mixture contained 0.08 mg./ml. of Azotobacter enzyme plus the following, in micromoles per milliliter: Tris buffer (pH 8.2), 150; MgClz , 10; ethylenediamine tetraacetate, 0.4; GDP, 13.7, and pApApA, 2. Aliquots were removed at different time intervals for quantitative paper chromatography. No polymer could be detected at 20 and 40 min; it appeared after 60 min. and increased progressively to reach a value equivalent to 8 pmoles of base per milliliter after 6J4 hours. Temperature, 37' C. monophosphate residues takes place, to give larger oligonucleotides and, finally, polymer. The polymer is nondialyzable against 0.001 M ethylenediamine tetraacetate, is precipitated by 2.5 per cent perchloric acid or 2 volumes of ethanol, and it does not migrate on paper chromatograms. Unfortunately, the amount of primer required for a reasonably rapid reaction with GDP is rather large. Thus, with 0.004 M pApApA the rate of polymerization of GDP is one fourth as fast as with ADP, and with less primer it falls off sharply. With limited amounts of enzyme it then becomes difficult to add more than about 9 guanosine monophosphate residues, on the average, per unit of primer and to do this on a reasonable scale. In contrast to the situation with ADP and UDP, oligonucleotides such as ApApUp do not stimulate polymerization of GDP. Polymers are also inac- 640 Annals New York Academy of Sciences tive, either because of inhibitory interactions or, perhaps, because the available concentration of terminal nucleoside residues onto which a guanylic acid residue can add is too low. In the reverse reaction catalyzed by this enzyme, namely, the phosphorolysis of polynucleotides to form nucleoside diphosphates, poly A and poly U are rapidly attacked by polynucleotide phosphorylase?, Substantial rates of phosphorolysis have been found for tobacco mosaic virus RNA, turnip yellow mosaic virus RNA, and highly polymerized yeast RNA.l0 On the other hand, commercial yeast RNA, which has been treated with alkali, is phosphorolyzed very s10wly.~ Commercial yeast RNA is considered to be made up of relatively short chains, many of them terminated by 3'-phosphomonoester and 2' ,3'-cyclic phosphoryl end groups." One possible explanation for its poor rate of phosphorolysis would be the resistance offered by such end groups. According to Singer,* oligonucleo- tides as large as a pentanucleotide are not phosphorolyzed if they possess the types of end group just mentioned. By contrast, oligonucleotides with a 5`- phosphomonoester end group are rapidly attacked until the molecule is reduced in size to a compound with 2 nucleoside residues. Conceivably, then, commer- cial RNA is resistant to enzymatic splitting because most of the chains have an unfavorable end-group structure. However, Grunberg-Manago has found that phosphorolysis of commercial RNA proceeds to the extent of at least 90 per cent.13 Possibly in a molecule larger than a pentanucleotide, enzymatic break down is possible even with a 3`-phosphornonoester or 2`,3`-cyclic phos- phoryl end group. In a preparation of RNA consisting of chains with an unsubstituted terminal nucleoside residue mixed with other chains in which a phosphate group is monoesterified at C-3' of the terminal residue, would chains of the second type inhibit the rate of phosphorolysis of chains that did not have such terminal phosphate groups? With this question in mind, Singer et a1.I4 recently meas- ured the rate of phosphorolysis of "5'-ended" oligonucleotides in the presence of "3'-ended" oligonucleotides. The results are shown in TABLE 2. It is apparent that ApApUp inhibits the rate of phosphorolysis of pApApA, but the effect is not great, even though a substantial amount of the `(3'-ended" oligo- nucleotide is present. Very little information is available on tbe end-group structure of RNA preparations, but the possibility should be kept in mind that inhibitory effects of this kind may be operating. A second reason for a slow rate of phosphorolysis of polynucleotides is the formation of multi-stranded chains. The interaction of poly A and poly U was first observed by Warner,15 and various aspects of this subject are discussed elsewhere in this monograph. Ochoa observed that when poly A and poly U were mixed in the ratio of 1: 1 the rate of phosphorolysis was considerably de- pressed, as compared with the rate for either polymer by itself.'O Grunberg- Manago made a similar observation and also noted that phosphorolysis was suppressed almost completely with a ratio of poly A to poly U of 1 : 2; under these conditions, a triple-stranded chain is formed.16 Her results with poly I are also of interest. This polymer is known, from the work of Rich," to exist as a random coil in dilute salt solution and as a triple helix in 0.6 M KC1. It was noted that phosphorolysis of poly I proceeded readily in dilute salt Heppel et ul. : Action of Polynucleotide Phosphorylase 641 and was suppressed nearly completely in 0.6 M KCl, whereas poly A was equally reactive in both concentrations of salt. At this point it is profitable to discuss a third reaction catalyzed by poly- nucleotide phosphorylase, namely the exchange of P82 and nucleoside djphos- phate. The exact mechanism of this interesting reaction is unproved. An explanation favored by Ochoa is that the incorporation of P? into nucleoside diphosphate results from synthesis of a small amount of polynucleotide, fol- lowed by its phosphorolysis. This is a reasonable supposition and suggests the possibility that oligonucleotides might stimulate the rate of the exchange reac- tion. Recently, Singer et aZ.l4 studied the effect of pApA and pApApA on the rate of exchange of inorganic P32 with ADP and UDP; a significant stimulation was PAPAPA (1.3) Poly A (0.9) PAPAPAPA (0.8) TABLE 2 PHOSPHOROLYSIS OF OLIGORIBONUCLEOTIDES* Substratet Additiont 1 Rate of phosphorolysis: I ApApUp (1.1) 14.5 APAPUP (1.1) 6.6 ApApUp (1.1) 1 55.7 PAPAPA (1.3) PAPAPAPA (0.8) poly A (0.9) * The reaction mixture contained, in 0.125 ml.: Tris buffer (pH 8.2), 5 pmoles; MgClz , 0.5 pmole; Py, 3.05 pmoles containing 448,000 cpm (polynucleotides as indicated), and 0.002 mg. E. coli polynucleotide phosphorylase (first ethanol step5). After 1 hour the reactions were stopped with perchloric acid, the nucleotides adsorbed onto charcoal, and the charcoal washedfreeof PIz and suspended in ethanolic NH, . Aliquots of thesuspension were plated and counted. t The numbers in parentheses are micromoles of polynucleotide per milliliter; for poly A this is expressed as adenine residues. $ Rates of phosphorolysis are expressed as micromoles of PIz incorporated into nucleotides per hour per milligram of enzyme. obtained (TABLE 3). These results agree with similar data obtained by Mii and Ochoa.'* The experiments shown in TABLE 3 were carried out with a prepara- tion of Azotobacter enzyme provided by Ochoa, which catalyzes the polymeriza- tion of ADP and UDP only after a lag period. This lag period is overcome by concentrations of pApA and pApApA similar to those described here. The maximum stimulation observed was not large, amounting to 140 per cent. It was hoped that more striking results might be obtained with GDP, since its polymerization shows an absolute requirement for an oligonucleotide primer. By suitable adjustment of the concentration of MgCl2 and the GDP-Pi ratio it was possible to obtain a rate of P!2-GDP exchange comparable to exchange rates for ADP and UDP* (TABLE 4). Under these conditions the rate of Pf2-GDP exchangewas stimulated more than threefold by 5 x lW3 M pApApA. A control experiment with pApApA, but no GDP, showed a very small incorpo- ration of Pf2 into nucleotide material; a correction for this was applied. * A detailed account of this is to be published. 642 Nucleotide added7 ADP ADP and pApA, 3 X ADP and pApApA, 5 X low4 M M UDP UDP and pApA, 3 X lWs M UDP and pApApA, 5 X lW4 M PAPAPA Annals New York Academy of Sciences Micromoles P: incorporated into charcoal-adsorbable nucleotides per hour, per miliigram Stimulation per cent 71 131 85 176 140 37 51 38 127 135 6 Substrate GDP ADP UDP Micromoles Pi" incorporated into charcoal-adsorbable nucleotides, per hour per milligram protein 45 51 49 Heppel et ul. : Action of Polynucleotide Phosphorylase 643 overcome by poly A and poly U, respectively. However, if the opposite pairs are used (for example, ADP and poly U), then the polymerization reaction is actually inhibited. Recently, Singer et d.I4 observed somewhat similar effects in studying the exchange reaction. Thus, when poly A, ADP, and P;' are present in a single reaction mixture, the incorporation of PQ2 happens to be equal to the sum of the ADP-P? exchange and the phosphorolysis of poly A, as measured separately. Similar data were obtained for the combination of UDP, poly U, and Pi ; however, when poly A is added to UDP or poly U to ADP, both exchange and phosphorolysis are inhibited. No such inhibition is ELUTION VOLUME, ML. FIGURE 4. Separation of oligoribonucleotides on an "Ecteola" column. Poly AU was made by incubating equimolar quantities of ADP and UDP with polynucleotide phosphoryl- ase; 30 mg. of the polymer was then digested with pancreatic ribonuclease and placed on a 44 X 1-cm. column of the modified cellulose adsorbent, Ecteola. Gradient elution was used, with 750 ml. of 0.015 M lithium acetate (pH 5.5) in the mixing chamber and 750 ml. of 0.4 M lithium chloride-0.015 M lithium acetate in the reservoir. Flow rate was 1.3 ml./ min. The peaks represent a homologous series of oligonucleotides; thus, Peak 2 is ApUp, Peak 3 is ApApUp, Peak 4 is ApApApUp, and so on. observed with GDP in the presence of either poly A or poly U. This is also true for thymine riboside pyrophosphate. Brief reference should be made to methods used in the separation of the several homologous series of oligonucleotides, which have been so useful in studies of polynucleotide phosphorylase. In the past, they have been isolated by paper chromatography12 and on Dowex 1-2x columns.1g Recently, in Rho- rana's laboratory, good separations of these oligoribonucleotides were achieved by Tener and Heppel (unpublished data) using modified "Ecteola" columns.2° Eluting conditions were similar to those previously employed* for oligodesoxy- ribonucleotides. FIGURE 4 shows the elution diagram obtained with the homologous series beginning ApUp, ApApUp. A similar record was obtained with the series beginning pApA, pApApA. * See discussion in Tener et al., this monograph. 644 Annals New York Academy of Sciences References polynucleotides: polynucleotide phosphorylase. J. Am. Chem. SOC. 77: 3165. nucleotides. coZi. Federation Proc. 16: 302. GLASS, Eds. Johns Hopkins Press. Baltimore, Md. with an enzyme from Escherichia coli. J. Biol. Chem. 226: 1077. 1956. adenosine diphosphate. Nature. 177: 790. 27: 222, cleotide phosphorylase. polynucleotide phosphorylase. Biochem. Biophys. 69: 119. 52: 565. ase. J. Biol. Chem. 232: 211. 1957. 1. GRUNBERG-MANAGO, M. & S. OCEOA. Enzymatic synthesis and breakdown of 2. GRUNBERG-MANAGO, M., P. J. ORTIZ & S. OCHOA. 1956. Enzymic synthesis of poly- Biochim. et Biophys. Acta. 20: 269. 3. LITTAUER, U. 2. Polyribonucleotide synthesis with an enzyme from Escherichia 4. KORNBERG, A. 1957. W. D. MCELROY and B. 5. LITTAUER, U. 2. & A. KORNBERG. 1957. Reversible synthesis of polyribonucleotides 6. BEERS, R. F., JR. Enzymic synthesis and properties of a polynucleotide from 7. OLMSTED, P. S. 1958. ADP-polynucleotide phosphorylase. Biochim. et Biophys. Acta. 8. MII, S. & S. OCHOA. Polyribonucleotide synthesis with highly pur%ed polynu- 9. SINGER, M. F., L. A. HEPPEL & R. J. HILMOE. 1957. Oligonucleotides as primers for 10. OCHOA, S. 1957. Phosphorolysis of natural and synthetic ribopolynucleotides. Arch. 11. MARKHAM, R. & J. D. SMITH. 1952. The structure of ribonucleic acid. Biochem. J. 12. SINGER, M. F. Phosphorolysis of oligonucleotides by polynucleotide phosphoryl- 13. GRUNBERG-MANAGO, M. Unpublished. 14. SINGER, M. F., R. J. HILMOE & M. GRUNBERG-MANAGO. 15. WARNER, R. C. Studies on polynucleotides synthesized by polynucleotide phos- J. Biol. Chem. 229: 711. 16. FELSENFELD, G., D. DAVIES & A. RICE. Formation of a three-stranded poly- 17. RICH, A. 1958. The molecular structure of polyinosinic acid. Biochim. et Biophys. 18. MII. S. & S.OCHOA. 1958. Personal communication. 1955. 1956. In Chemical Basis of Heredity. : 579. 1957. Biochim. et Biophys. Acta. 26: 445. Biochim. et Biophys. Acta. 26: 447. 1958. Unpublished. phorylase. nucleotide molecule. J. Am. Chem. SOC. 79: 2023. Acta. 29: 502. 1957. 19. VO~KIN, E. & W. E. COHEN. 1953. On the structure of ribonucleic acids. J. Biol. 20. PETERSON, E.A. & H. A. SOBER. 1956. Cellulose ion-exchange adsorbents. J. Am. Chem. 206: 767. Chem. SOC. 78: 751.