Brown Hepeel and Hilmoe Nucleotides. Part XXI V.* The Action of Some Nucleases on Xirnple Esters of Nonoribonucleotides. By D. M. BROWN,L. A. HEPPEL,and R. J. HILMOE. [Reprint Order No. 4576.1 An enzymically active fraction from spleen converts adenosine-3’ benzyl phosphate and cytidine-3’ benzyl phosphate into adenosine-3’ phosphate (adenylic acid b) and cytidine-3’ phosphate (cytidylic acid b) respectively. The corresponding 2’-esters are not affected. The same fraction yields only the 3’(b) -mononucleotides from ribonucleic acids and polyribonucleotides. Taking these in conjunction with previous results on the specificity of ribo- nuclease (Brown and Todd J. 1953 2040) it is concluded that the inter- nucleotidic linkage in ribonucleic acids involves the 3’-and not the 2’-position in both purine and pyrimidine nucleotide residues i.e.that the ribonucleic acids are 3’ 5’-linked polynucleotides. Comparable observations with diesterases from other sources are described. RECENTwork on the structure of the ribonucleic acids has had as its objective the elucidation of the nature and position of the internucleotide linkage of the polynucleotide chain. From a consideration of the mechanism of chemical hydrolysis by acid and alkali it became clear that the nucleoside residues comprising the chain were linked through the phosphoryl group at C(3t) or C(zt) and C(5n)(e.g. I) (Brown and Todd J. 1952 52). In practice it has been necessary previously to refer to these positions as a 6 and C(5t)since * Part XXIII preceding paper.[19541 Nscleotides. Part XXI V. the orientation of the phosphoryl group in the isomeric monoribonucleotides derived from hydrolysates of ribonucleic acid had not been established. Recently however Khym Doherty -Volkin and Cohn (J. Amer. Chem. SOC.,1953 75 1262) have given chemical evidence that adenylic acid b is adenosine-3’ phosphate (I1; R = adenine residue R’ = H). Cytidylic acid b is considered to be cytidine-3’ phosphate (11; R = cytosine residue R’ = H) on physical grounds (Loring Hammell Levy and Bortner J. Biol. Chem. 1952 196 821; Cavalieri J. Amer. Chem. SOC.,1952 74 5804) and hence uridylic acid b is the 3’-isomer (11; R = uracil residue R’ = H) (Brown Dekker and Todd J. 1952 2715). Recent evidence from enzymic studies (Shuster and Kaplan J.Biol. Chem. 1953 201,535) confirms the view (Cohn J. Cell. Cow#. Physiol. 1951 Suppl. 1 38 21) that the b isomers of all four monoribonucleotides are alike in the position of the phosphoryl group in the ribofuranose residue. Any conclusions regarding the position of the internucleotidic linkage in the polynucleotides depend on the validity of the structures assigned to the mononucleotides and in the present paper we accept those indicated above. The decision between C(3tland C(2pj as a linkage point in the polynucleotides was reached in the case of the pyrimidine nucleoside residues by a study of the action of the pancreatic enzyme ribonuclease on some simple benzyl esters of the isomeric uridylic and cytidylic acids. Uridine-3’ and cytidine-3’ benzyl phosphate (11; R = uracil and cytosine residue respectively R’ = CH,Ph) were hydrolysed via the corresponding nucleoside-2’ :3’ phosphate to uridine-3’ and cytidine-3‘ phosphate respectively ; the 2’-esters (I11; R = uracil and cytosine residue respectively R’ = CH2Ph) were unaffected (Brown Dekker and Todd J.1952 2715; Brown and Todd J. 1953 2040). Since the mechanism of hydrolysis of these simple esters by the enzyme was entirely analogous to that observed for the ribonucleic acids (Markham and Smith Biochem. J. 1952 52 552 558) it was concluded that the pyrimidine nucleotide residues were linked at the 3’-position in the ribonucleic acids. The specificity of ribonuclease clearly did not permit any conclusion concerning the linkage of the purine nucleotide residues.Evidence on this was sought in the action of other related enzymes. There are numerous indications in the literature that nucleases with specificities different from that of ribonuclease exist in plant and animal tissues. E.g. Schmidt Cubiles and Thannhauser (J. Cell. Comp. Physiol. 1951 Suppl. 1 38 61) found that pancreas contains a fraction which hydrolyses further the limit polynucleotides produced by ribonuclease digestion of ribonucleic acids. Volkin and Cohn (Fed. Proc. 1952 11 303) showed that extracts of spleen prepared according to Maver and Greco (J.Biol. Chem. 1949 181 861) degrade ribonucleic acids to the 3’-mononucleotides. Fractionation of spleen has led to three distinct enzymically active fractions (Heppel and Hilmoe Fed.Proc. 1953 12 217) of which that designated Fraction I11 has been used in the present work ; degradation of ribonucleic acids and some ribonuclease-resistant polynucleotides by Fraction I11 has been studied (Heppel Markham and Hilmoe Nature 1953 171 1152) concurrently. It was found that ribonucleic acids and certain polynucleotides are split by the enzyme preparation to mononucleotides of which the purine representatives were adenosine-3’ phosphate and guanosine-3’ phosphate. No evidence of intermediate nucleoside-2’ 3’ phosphates was observed. It was concluded that the diesterase split the internucleotide linkages without the possibility of migration of the phosphoryl group so that the purine nucleoside residues in the intact polynucleotide had been linked at the 3’-position.This although plausible would be rendered more certain by a study of the action of the enzyme on isomeric nucleotide esters in which the phosphoryl group is known to occupy the 2’- and the 3’-position of the nucleoside residue. Brown Heppel and Hilmoe We have now studied the action of the spleen fraction on benzyl esters of adenosine-2’ and -3’ phosphate and cytidine-2’ and -3’ phosphate following the reactions on paper chromatograms. The benzyl esters of the two adenylic acids as their mixe-d barium salts have previously been described (Brown and Todd J. 1952,44); they have now been separated by ion-exchange chromatography and the free acids obtained in crystalline form. The cytidine derivatives have already been described (Brown and Todd J.1953 2040). The adenosine-3’ and cytidine-3’ benzyl phosphates when treated with the spleen fraction were completely converted into adenosine-3’ phosphate and cytidine-3’ phosphate ; no evidence for a cyclic phosphate intermediate was found on paper chromatograms. The 2’-isomers mere however unaffected even though for adenosine-2’ benzyl phosphate the concentration was varied 80-fold. This was not due to enzyme inhibition because incubation of a mixture of the 2’- and the 3’-ester resulted in approximately complete splitting of the latter. A stricter comparison of the two esters was made by using nearly comparable quantities of substrate and varying the concentration of enzyme widely (Table 1). This showed that hydrolysis of the 2’-isomer if it occurred must be at a rate less than 1/3000 that of the 3’-isomer.TABLE1. Action of enzyme fractions from spleen intestinal nmcosa fiotato and .rye grass on certain benzyl esters. Incubation Enzyme solution (ml. per Substrate (benzyl Concn. of substrate (,umoles/ml.) Hydrolysis (hr.) ml. incubation mixture) phosphates) initial final change (%I (a) Spleen fmcfion. 32 1 0.57 0.13 Cytidine-2‘Cytidine-3’ 1.4 3.8 1.4 3.0 0 0-8 0 20 24 1.0 Adenosine-2’ 2.5 2-5 0 0 32 32 29 0.57 0.57 0-57 1) as 4-13 0.6 0.06 4.8 0.6 0.06 0 0 0 0 0 0 1 0.13 Aden&ne-3’ 9-4 4.4 5.0 55 1 1 1 1 0.13 0,065 0.03 1 0.016 , , 2.4 7.6 2.9 1.9 0.4 4.5 2.3 1.7 2.0 2.8 0.6 0.2 80 35 20 11 7 8 0.002 0.001 , sD 3.3 7.7 3-1 7-6 0.2 0.2 6 3 (b) Intes2in.al nzucosa fraction.32 7 0.65 0.65 Cytidine-2’Cytidine-3’ 1.9 4.8 1-9 3.9 0 0.9 0 20 32 0.65 Adenosine-2’ 1.7 1.7 0 0 7 0.65 Adenosine-3’ 3.9 2.5 1.4 35 (c) Rye grass fraction. 24 1 0.75 0-75 Cytidine-2’Cytidine-3’ 1.9 4.4 1.9 0.4 0 4.0 0 90 30 1 0-75 0.75 Adenosine-2‘ Adenosine-3‘ 2.7 6.1 2.7 0.1 0 6.0 0 95 1 0.07 9 5.7 0-7 5.0 85 (d) Potato extract. 28 4 0.6 0.8 Cytidine-2’Cytidine-3’ 1-3 2.9 1.3 1.1 01.s 0 60 28 4 0.8 0.8 Adenosine-2‘ Adenosine-3’ 2.0 3.6 2.0 2-2 0 1.4 0 40 1 The same enzyme solutions were used throughout (see Experimental section). Although no evidence for the presence of nucleoside-2’ 3’ phosphates in the solution during the enzymic hydrolysis was found the enzyme preparation does split the cyclic phosphates under comparable conditions yielding mononucleotides * (cf.Heppel * [Added 24.9.53.1 Observations made since submission of this paper by one of us (L. A. H.) with Mr. P. R. Whitfeld and Dr. R. Markham show that this enzyme preparation yields only nucleoside-2’ phosphates from the corresponding cyclic phosphates. This does not affect the conclusions drawn from the present work. [1954] Nucleotides. Part XXI V. 43 Markham and Hilmoe loc. cit.) and Table 2 shows that the rates are of a similar order of magnitude. The mechanism of hydrolysis of the esters is of some importance since ribonuclease which has a closely related specificity acts by a mechanism essentially TABLE2. Action of enzyme fractions from spleen intestinal mucosa fiotato and rye grass on gunnosine-2' :3' fihosphate.Concn. of guanosine- Enzyme solution (ml. 2' 3' phosphate Hydro-Incubation per ml. incubation (pmoles /ml.) lysis (hr.1 mixture) Enzyme initial final change (%) 1 0.6 Rye grass fraction 2-2 1.1 1.1 50 1 0.6 Potato extract 2.9 2.2 0.7 25 1 0.09 Spleen fraction 2.7 2-23 0.4 15 1 0.65 Intestinal mucosa fraction 4.5 3.0 1.5 36 similar to that of chemical hydrolytic agents (Brown and Todd loc. cit.) in which cyclic intermediates are involved. Present evidence is insufficient to allow more detailed discussion. The same spleen fraction catalyses the nearly complete hydrolysis of ribonucleic acids well over 90% being converted into mononucleotides (Heppel Markham and Hilmoe Disappearance of adenosine-3' benzyl phosphate when incubated with spleen fraction at 37".Each ml. of incubation mixture contained 0.2 ml. of enzyme solution 16.3 pmoles of the ester and 0.3 ml. of 0.5~-potassiuwz phosphate buffer (pH 7.4'1. loc. cit.). Since then only the nucleoside-3' benzyl phosphates are hydrolysed by the enzyme it follows that the linkage point in the ribonucleic acids must also be at the 3'-position of the nucleoside residues regardless of the mechanism of the enzymic hydrolysis. Recently it has been shown by Whitfeld and Markham (Natztre 1953 171 1151) that only purine nucleoside-3' phosphates are formed when dinucleotides derived from ribo- nuclease hydrolysates of yeast ribonucleic acid are subjected to a novel chemical degrad- ation. The method (Brown Fried and Todd Chem.and Ind. 1953 352) cannot involve phosphoryl migration. Thus if one accepts the structures assigned to the isomeric monoribonucleotides and the accumulated evidence that C(51 is involved as one of the linkage points of the nucleoside residues the ribonucleic acids must be considered to be polynucleotides in which the individual nucleoside residues are linked by phospho-diester groupings at the Cc3R) and the C(,.) positions as in (I). Chain branching if it occurs either by incorporation of phospho- triester groupings or at C(zr>of the sugar residues of the main chain (cf. Cohn Doherty and Volkin " Phosphorus Metabolism," Johns Hopkins Univ. Press Baltimore 1952 Vol. 11 p. 339) is not discussed here since the present experiments do not bear on the point.Several other enzyme preparations which catalyse an extensive degradation of ribo- nucleic acids have been tested for diesterase activity. A fraction from intestinal mucosa (Heppel and Hilmoe loc. cit.) a dialysed potato extract and a relatively crude fraction from rye grass (Shuster and Kaplan loc. cit.) rapidly hydrolysed the 3'-esters but had no Brown Heppel and Hilmoe effect on the 2’-isomers (Table 1). This gives further support to the conclusions regarding the structural features of the ribonucleic acids reached in the spleen nuclease experiments. The intestinal mucosa fraction like the spleen preparation converts adenosine-3’ benzyl phosphate into adenosine-3’ phosphate ; however the plant extracts cause rapid conversion to the corresponding nucleoside.In the case of potato A. Kornberg (personal communication) has shown that during extensive purification the ratio of diesterase to non-specific phosphomonoesterase remained constant. Unless further work results in a dissociation of these activities we have no simple way of determining the nature of the first reaction product with this diesterase preparation. The rye grass fraction kindly supplied by Drs. Shuster and Kaplan converts the nucleoside-3’ benzyl esters into nucleosides under conditions where nucleoside-2’ phosphates would not be attacked. This fraction possesses monoesterase activity specific for 3’-nucleotides (Shuster and Kaplan Zoc. cit.). It is probable that the conversion of the benzyl esters into nucleosides by this preparation involved the intermediate formation of nucleoside-3‘ phosphate i.e.that the initial hydrolytic step was analogous to that catalysed by the spleen and the intestinal fraction. Kinetic studies were not undertaken since the 2’-esters were not detectably split while the 3’-esters were completely hydrolysed. However the Figure suggests that the hydrolysis of adenosine-3’ benzyl phosphate by the spleen fraction follows first-order kinetics. It was felt desirable to relate the activity of these four enzyme preparations towards the benzyl esters with their activity against a natural substrate. Table 2 indicates that the rates of hydrolysis of guanosine-2’ 3’ phosphate are of a similar order of magnitude. In the present investigation we have found that fractions from spleen potato and intestinal mucosa hydrolyse adenosine-3’ benzyl phosphate but not adenosine-5’ benzyl phosphate.Preliminary experiments with another diesterase derived from intestinal mucosa (Fraction V; Heppel and Hilmoe Zoc. cit.) show the opposite specificity; the 3’-ester is not hydrolysed but adenosine-5’ benzyl phosphate is attacked. Thus it can be seen that with the synthetic alkyl esters one can obtain new information concerning the structural groups required for the specificity of diesterases. This can be used in following their separation from one another and for classification. EXPERIMENTAL Purification of Enzyme Fractions.-(a) From spleen and intestinal mucosa. Further puri- fication is being continued by one of us (R.J. H.) and a detailed description will be published later. For our experiments the procedure was as follows. All operations were at 3” except as noted. Calf spleen was homogenized with three volumes of cold O.S5~-sucrose and the mixture adjusted to pH 5.1 with acetic acid. A bulky precipitate formed which was collected by centrifugation and washed with cold ( -10’) acetone. The dried powder was extracted with 20 parts (v/w) of O.B~-acetate buffer (pH 6) and the solution fractionated with ammonium sulphate first at pH 4.9 then at pH 8 each time selecting broad fractions containing most of the activity. The enzyme solution used in these experiments was dialysed against distilled water and contained 7 mg. of protein per ml. Calf intestinal mucosa was homogenized with 3 volumes of cold 0.25~-sucrose adjusted to pH 5.1 with acetic acid and all insoluble material removed by centrifugation.The supernatant solution was twice fractionated with ammonium sulphate at pH 4.6. A fraction precipitated between 0-7 and 0.85 saturation was dialysed against distilled water and constituted the enzyme solution used for this work. It contained 9 mg. of protein per ml. and was free from phosphomonoesterase activity. (b) From potato extract. Skinned potatoes were homogenized in a Waring blender with one- half of their weight of distilled water then filtered through paper and the filtrate was saturated with ammonium sulphate. The precipitated proteins were collected by centrifugation and dialysed against cold distilled water to give a solution containing 3.0 mg.of protein per ml. (c) From rye grass fraction. This was obtained from Drs. Shuster and Kaplan and represented the stage of 2-fold purification described by them (cf. J. Bid. Chem. 1953 201 635). The solution used contained 1.5 mg. of protein per ml. Part XXI V. [19541 Nucleotzdes Adenosine BenzyZ Phosphates.-Yeast adenylic acid (0.5 g.) was treated with phenyldi- azomethane (from 2-3 g. of benzaldehyde hydrazone) in dimethylformamide (5 ml.) and kept for some time. Ether water and barium carbonate were then added. The solution was shaken for several hours with changes of ether. The aqueous phase was separated filtered through Hyflo Supercel and evaporated under reduced pressure.The product was twice reprecipitated from water (3-4 ml.) by acetone and dried (0-44 g.). This material was dissolved in water (50 ml.) brought to pH 8 and run on to a column (9 cm.2 x 11 cm.) of Dowex-2 resin in the formate form. After washing with water (500 ml.) elution was commenced with 0-1N-formic acid. Mononucleotides were removed with the first 5 1. of solvent which were discarded. Adenosine-2’ benzyl phosphate was slowly removed from the column with the next 2 l. and the remainder rapidly removed by changing to 0.5~-formic acid. Continued elution with this solvent then removed adenosine-3’ benzyl phosphate. The course of fractionation was followed by observing the optical densities of the fractions (ca. 20 ml.) at 260 mp. The fractions containing adenosine-2’ benzyl phosphate were evaporated at 20 mm.and the residue crystallized from water. The product formed small irregular prisms from water which retained water tenaciously (Found in material dried at 105°/0-1 mm. C 45.1 ; H 4.5; N 15.0; P 7-8. C,,H,,O,N,P,H,O requires C 44-8; H 4-8; N 15.3; P 6.8%). The fractions containing adenosine-3’ benzyl phosphate were evaporated at 14 mm. at room temperature with additions of water at the later stages to reduce the formic acid concentration. The residual oil crystallized from water in rosettes of small needles (Found in material dried at 105°/0.1 mm. C 43.2; H 5.0; N 14-5; P 6.7. C1,H,,0,N,P,2H,0 requires C 43.1 ; H 5.1; N 14.7; P 6.6%). The two substances gave single spots on paper chromatograms with butyl alcohol-acetic acid-water (RF0.42 and 0.47) corresponding to adenosine benzyl phosphate a and b respectively (Brown and Todd J.1952 44). Guanosine-2 3’ Phosphate.-This compound was obtained (Markham and Smith Zoc. cit.) by heating dialysed ribonuclease digests of yeast ribonucleic acid with excess of barium carbonate at 100” for 1 hr. followed by chromatographic separation. Incubation of Enzyme Fractions with the Benzyl Esters and with Guanosine-2’ :3’ Phosphate.- All incubations were carried out at 37”. For experiments with spleen fractions the total volume was 0.07 ml. ; this contained amounts of enzyme solution and concentrations of substrate as indicated in the Tables and 0.015 ml. of 0.5~-potassium phosphate buffer of pH 7.4. Incub-ations with intestinal mucosa and rye grass fractions were similar except that the total volume was 0.06 ml.For experiments with potato extract the total volume was 0-05 mi. and the mixture contained 0-01 ml. of M-ammonium acetate buffer of pH 5. Control incubations were carried out both with omission of enzyme and omission of substrate. It was also deterinined that the enzyme preparations remained active at 37” for at least 2 days. In preliminary experiments it was found that the enzyme fractions completely hydrolysed the 3’-esters under conditions where no detectable splitting of the 2’-esters took place. The concentration of ester was 2-5 pmoles per ml. and the formation of 0.02 pmole of product per ml. could have been readily detected. Quantitative experiments were then carried out as follows.At the end of incubation an aliquot portion was applied as a spot or thin line on Whatman No. 3 MM paper. Further reaction was halted by encircling the spot with acetone and allowing this to diffuse into the regions where applications had been made. After chromatography the sheets were dried and photographed in ultra-violet light (Markham and Smith Biochem. J. 1949 45 294; 1951 49 401). The ultra-violet-absorbing regions of the paper were cut out together with appropriate control regions. Elution was at room temperature for 20 hr. with 5 ml. of 0-1N-hydrochloric acid after which the absorption at 260 m.p was measured with the Beckman model DU quartz prism spectrophotometer. The results are shown in Tables 1 and 2. For the paper chromatography of the compounds discussed in this paper several solvent systems were used (Whatman No.3 MM paper solvent descending). Good separation of the benzyl esters from nucleotides or nucleosides was obtained with the system isopropanol-water (7 3; v/v),with the further addition of 0-35 ml. of aqueous ammonia (d 04380) per 1. of gas space in the tank (poured into the bottom of the tank) (cf. Markham and Smith Biochem. J. 1952 52 552). The adenylic acid formed from its benzyl ester by fractions from intestine and spleen was identified as the 3’-isomer by migration with the solvent system saturated aqueous ammonium sulphate-M-sodium acetate-isopropanol (80 18 2 ; v/v) (Markham and Smith Zoc. cit.) by comparison with an authentic specimen of adenosine-3’ phosphate.The nucleotide formed from cytidine-3’ benzyl phosphate by the spleen fraction was shown Christie Kenfler and Todd to be entirely cytidine-3’ phosphate by means of ion-exchange chromatography (Brown Dekker and Todd ZOC. cit.). One of us (L. A. H.) participated in this work as a member of the National Institutes of Health Bethesda temporarily assigned to the Molten0 Institute. Thanks are offered to the John Simon Guggenheim Memorial Foundation for a Fellowship (to L. A. H.) to the American Cancer Society for a travel aid grant (to L. A. H.) to the Rockefeller Foundation for financial assistance (to D. M. B.) and to Professor A. R. Todd for his continued interest. UNIVERSITY CAMBRIDGE. CHEMICALLABORATORY, PLANTVIRUS RESEARCH CAMBRIDGE. UNIT,MOLTENOINSTITUTE NATIONAL OF INSTITUTES HEALTH,BETHESDA. [Received August llth 1953.1